Ummm hopefully this new top of the line SKU will be able to give the advetised single core turbo and all cores turbo and definitively not on only 5% of the stockpile of chips. Yes because if the crap story will repeat itself, nearly all the reviews on selected SKUs for press from AMD will have zero relevance, but it will be only a good marketing show. IMO the delay is for a more accurate binning at TSMC.
Secondly, Intel mentions Max single core turbo also, but more as a guideline. Intel's chips never, under any circumstance, reach max turbo at their rated TDP. (Thermal or Power budget). Turbo speeds has always been "If power and cooling allow", even on Intel... especially on Intel too. And fun fact, usually to get the advertised turbo speeds on Intel, you can't be using any AVX, else you hit the offset.
There's only 1 guarantee, the base clocks. But who cares about a Ghz number, when what should matter is actual performance? AMD's FX9590 was 5ghz all core, but was still a crap CPU.
>Intel's chips never, under any circumstance, reach max turbo at their rated TDP.
What's with AMD fanboys and spreading miainformation. Silicon lottery is a thing, there's dies out there that can turbo at base TDP. And most dual core Skylake U series CPUs will run at turbo indefinitely due to the fact their turbo clock is so low.
>Turbo speeds has always been "If power and cooling allow", even on Intel... especially on Intel too.
No, all Intel CPUs will turbo to some extent. The amount of time they spend in turbo is what changes.
Jorgp2 too bad intels chips dont run at base TDP under use.i guess you havent read this : https://www.anandtech.com/show/13544/why-intel-pro... if you do limit intels cpus to their rated TDP, their performace, goes down the toliet. the 9900K for example, TDP of 95 watts, actual usage, 150-200 watts " No, all Intel CPUs will turbo to some extent. The amount of time they spend in turbo is what changes. " ONLY if the cpu has adequate cooling.
i'll chip in and say that it's you. you will most likely not see any chips that can go full turbo at TDP in 2019. The difference between TDP and peak load are too far from each other and while the scilicon lottery is a thing, in most cases intel will let an amazin i5 upgrade to a low tier i7 and a low tier i7 upgrade to a high tier i7. they will push it further if it comes out as a great chip, so the only case where this would normally be seen would be on a top tier i7 that has just hit the golden everything, and it's too rare to talk about for most people.
What's with Intel Fanboys instantly calling people AMD fanboys when they're simply stating things about Intel CPUs that are facts?
And no, you're wrong on that front too. anything not a Core i5 or i7 doesn't have Turbo tech built in... unless I'm mistaken, I haven't purchased or seen a Core i3 since the Coffee Lake 8000 series...
Turbo speeds are never a guaranteed thing, Intel admits this. I'm no fanboy, I currently own no AMD devices, and I've got 4 computers to my name (Core i7 4790k and Xeon E3-1245-V5 make up the two most important PCs I own)
You clearly don't understand the difference between what a chip can do, and what the TDP is. Going above base speeds will take more power, but how a chip is rated when it comes to TDP is very different between Intel and AMD.
Intel ONLY reports the TDP for base speed. So if the i9-9900k says 95W TDP, that is at the base speed of 3.6GHz. The fact that the i9-9900ks has a TDP of 127W with a base speed of 4.0GHz indicates very clearly that just going from 3.6GHz to 4.0GHz boosts the TDP from 95W to 127W. So that magic 5.0GHz that people want isn't going to be at 95W, or 127W, it's going to be upwards of 200W.
AMD rates TDP for typical usage, and while that is a bit vague, it means that a 105W TDP CPU will be enough to run above base speeds and stay within the official TDP rating.
Again, Turbo/Boost speeds and the TDP rating are two different things. If power delivery to the CPU were limited to official numbers, you wouldn't see any Intel chips getting above 4.6-4.7GHz even with golden samples. Throw unlimited power at the chip, and you get higher clock speeds.
I think you should read the article that refers to how AMD's turbo's definition and how theirs works vs Intel, as its quite enlightening and I do agree with the observations made by Anandtech. I think AMD's done a great job and its nice to know that the CPU you're buying from them is pretty much at max performance, and could probably undervolt for better performance instead of overvolt and overclocking vs Intel's overclock for more performance because they're leaving more on the table for the user to exploit.
Hmm, how did we get from "low OC headroom is a drawback" (Intel) or "all chips should be able to OC" (Intel) to "0 OC headroom is good" (AMD)?
Also, OC is not an exploit, it's just a feature of almost all (AMD) or selected (Intel) SKUs. That's why you buy Intel K series - to Oc them.
Next thing - majority of AMD Ryzen 3000 CPUs don't hit Boost Clock ever. Watch De8auer video, hunders of people took part in survey and it turns out in some SKUs only few (<10) % of units hit Boost clock EVER.
I'm going to ask you to think about something. Who do you think is more likely to take part in a survey? Someone who is pissed off and has something to say? Or someone who is completely happy with what they have and have stopped thinking about it? It's a fact that voluntary surveys are skewed negative because of people are more like to complain if something isn't right than write about how good it is. I wouldn't take survey numbers like they're a scientific study.
There's that bias, but I also think Der Ba8er's survey has a problem, because of Windows Scheduler. AMD only guarantees the boost clock on one core, the one marked with the gold star in Ryzen Master. When I run the CB single core test on my Ryzen 3600 it does not get scheduled to that core, but to a seemingly random core. In fact, in spite of the name, I see loads peak on different threads during the run and I am pretty sure not all are on the same core (I use CB20 instead of CB15.) And there is no way to easily force Windows to use the gold star core (Maybe it can be done with Process Lasso?)
If this is the case, it is no wonder that in Der Ba8er's survey the number of Ryzen 3000 CPU's that can't reach their advertised boost clock gets worse with a rising number of cores. The chance that the benchmark is run on the gold star core just gets lower.
Of course, this is not the entire story, because AMD's ABBA Agesa revision did help. But I still think Windows not assigning the CB single core benchmark run to the right, 'gold star' core plays an important role in Der Ba8er's survey results.
tabalan... has de8auer ever talked about how much power intel uses to reach the clocks and performance it gets ? like teldar said.. those surveys are bias towards the negative. and should be taken as biased
So it's clear from your reply that you DIDN'T read the article, because the impossibility of measuring an instantaneous boost without tainting the result is covered pretty thoroughly.
Next Thing is already here: with bioses built around AGESA ABBA, all Zen 2 CPUs are once again hitting their max advertised boosts MHz with ease--just as they did with with the early 1002 AGESAs--just like mine. How did we get to "overclocking is a guaranteed function of all cpus," ? when the *the fact* is just the opposite--overclocking is never guaranteed in any form by any CPU manufacturer. Oh--that includes Intel--which I guess I have to spell out. "OC headroom" is not something smart people care about--CPU performance, however, is very much something people care about. MHz, remember, is only half the performance picture--the other half--the more important part--is IPC execution. OC is not a "feature" guaranteed or warrantied" by any CPU manufacturer to any degree whatsoever. Sorry that you don't understand that OC is not a "feature"....*snicker*. You might be interested to know that 90% of the people who buy CPUs made by anyone--at least 90%-- do not bother OC'ing them at all.
I think the biggest thing in this change is that we shouldn't have to overclock. With Intel, we've gotten used to them lowballing us and nerfing the chips to force us to overclock in order to get what we paid for, and people even made that into a sport. Overclocking became a 'feature' because Intel has been artificially sandbagging and cheating us for years.
With the Zen chips, it is more of "you get what you paid for out of the box" kind of mentality and that different tiers of products actually offer valid upgrades for the extra money instead of just lesser degrees of artificial nerfing from the manufacturer.
I'm neutral on this. On the one hand, AMD chips not meeting their boost and having low head room is a drawback, however the performance we are seeing in the benchmarks is with those same lower boosts and lower headroom.
Yes, it's a downer that they don't seem to hit the peak speeds, but at the same time, if you have a multi threaded load, they are still out performing Intel by a serious margins and nearly on par in single threaded loads. That is ultimately what matters.
The demand for AMD's Zen2 has blown away all of the company's internal estimates--the demand for these fantastic CPUs exceeds AMD's studied projections. With demand skyrocketing, there is no way the original estimates could hold, because the nice thing about this from AMD's perspective is that their most expensive Zen2 CPUs are the most in demand! Seems to be a trend here.
Ryzen chips are boosting just fne. That being said, I read somewhere the other day that the new Threadripper chips (or at least 1 of the SKUs) is using a new socket. I wonder if Ian has any insights into this?
Glad to see them finally set a date. They've been saying Q4 for launch since the Q2 Financial call for TR3. I've been worried the retail side wouldn't see much because the cloud companies were sucking up all the silicon like the original Epyc that you couldn't buy for 6months after release because Google, Amazon and the others were soaking up the whole production run. I hope the same thing doesn't happen with TR3 and AMD has stockpiled enough to get it into the retail channel as well.
You still can't really buy the 3900X on Amazon at the MSRP or below unless you get lucky and login right when a shipment arrives. I've only seen it available once at the MSRP. So AMD still has really strong demand for Ryzen Gen3 in the retail channel. I expect the 3950 will be even worse and the TR3 could suffer the same fate if they don't stockpile before release to retail.
I agree. Having a product which is good enough for demand to strain supply is a good thing for AMD - especially if it's the very profitable EPYC line - just don't wait so long to fill retail channels that competition has time to catch up, or for consumer interest to wane.
This doesn't look good for future Ryzen APU's. Since there is already an issue with delivery times on TSMC's 7nm process for orders that have been placed months ago it will just get worse with all the mobile SoC (Qualcomm's 865 and Apple's A13) required over the next few months. It could take another year before we see any Ryzen 3 based APU's on the market.
Ryzen 3 APUs are out now. They always fab the APU chips on the previous process with the previous core design. Ryzen 4 will be the first 7nm Zen 2 APUs.
The bulldozer APUs came out alongside the desktop CPUs with the same core generations, IIRC. Heck, the bulldozer cores got an extra iteration on mobile that the desktop never saw, as did K10 with the llano APUs.
First thing I should say is that I am An AMD fan & every time they release a new gen of Ryzen always hope that they do well.
No on to the part some won't like most likely. I do like the idea of the chiplet it makes tings more cost effective and just easier to make the CPU than going with a huge core setup like the other guys are doing. The issue I have is that when you have a CPU like the 3900x & 3950x that use multi chiplets. AMD needs to make sure that each chiplet is binned in such a way that at any point any core (single core Turbo) can reach the advertised speed for turbo boost.
From what I am seeing here is you may get one chiplet that has a few better cores in it but the other chiplet only has cores that will only get to just below the advertised turbo boost. They need to have binning that has both chiplets have cores in it that can do the max turbo. I am not saying oh they have to be able to full turbo on all cores at once what I am saying is that any core at any given time needs to be able to run at the max turbo speed not just a few select cores once and a while.
Heck even on my old i7 2600K all 4 cores can reach the max turbo with ease and those same 4 cores can reach up to 5.1GHz and this CPU is from what 2011-2012 time frame and made on 32nm process. No max turbo is not guaranteed just base clock is but when they advertise clock speeds and turbos those CPU's better be able to do them even if it is only for a few seconds at a time and every core should be able to do it not just a couple cores in the CPU now and then. A said I am talking about single core turbo not whole chip speed.
This is one of the reasons I am still on my Sandy bridge is the rather lack luster MHz on the Ryzen platform. Yes they perform well even at the lower clock rates. The other reason is that on the Intel side they priced themselves out of my wallet a long time ago oh and the ever changing socket game they like to play as well pretty much made me not want to have Intel in my system at least for now. This is just my opinion nothing more and nothing less.
I don't get your logic? AMD are only saying that turbo boost applies to a minimum of one core and even then no guarantee on how long that turbo will be for. As long as one core momentarily hits the advertised box speed then that is it. It doesn't matter which core or which chiplet so all AMD needs to make sure is one chiplet contains 1 core that can hit that turbo speed. I really think all these AMD/Intel fan boys have lost the plot and would be better off taking up another hobby such as drag racing. Pretty sad when humanity has come to the point where we can take a pile of sand though unimaginable technological and industrial processes just so some kid in his basement can run a stupid benchmark and then complain about a few missing mhz or a couple missing frames in a game..
This is not charity, your comment sound like 'Wah Wah wah, dont say bad things about AMD' AMD is another greedy corp that wants to make money and we as Consumers that pay their bills should get the maximum, without if or maybes. AMDs turbo is pure scam, they can easily write on the box 5Ghz and 6Ghz, as long as it hits the score for one nano second its OK. The turbo should be useful not useless, the way it is now, AMDs turbo is advertising scam.
Amd specifies a single core turbo of 4.6Ghz for 3900x and that is reached. They never said that all core turbo is also 4.6Ghz. And nevertheless, it is not all about frequency. Even with this deficit 3700x and 9900k are even in compute tasks
rocky12345 and the point is about your 2600k ?? so what if it can reach 5.1 ghz, almost any equivalent newer chip then that, would more then likely out perform it, even if clocked lower. what speed a cpu runs at, doesnt mean much, its the IPC the cpu does that is important, as for the comment on the turbo thing, did you read this : https://www.anandtech.com/show/14873/reaching-for-... ?????
hmm an AMD fan who has a intel chip and bashes AMD? it has never been promised or stated that all cores could attain the max some do some dont, if you want real dishonesty look at the other team who claim TDPs that are frankly pure fantasy, I think we are also getting the result as stated on a few articles of the differing interpretation of the two big players when it comes to turbo Intel jacks up the power and gives all cores AMD stays within the TDP but only hits that max momentarily because the chip is more efficient to start with and simply works better at a given TDP if you restrict the chips to a given or stated TDP then see how well the Intel boosts it would paint a very different picture. FWIW the only CPU I have lost for no apparent reason was a team blue one a core 2 duo one, my one and only ever Intel purchase died after 2 years of modest use I was strictly AMD before or since, I just find them more honest and in daily use they have always felt snappier in a windows environment (which is probably purely in my head).
They probably didn't anticipate such a big success for Zen 2. They probably have the same situation on the server side with epyc so I am guessing they give priority to that which leaves consumer side with not much.
Also I believe that amd gets a small chunk of tsmc production capacity since it is a new customer. It is a good problem to have high demand, but if you can't satisfy it then it means you lose money.
Rather have a small delay than no stock. Current high end Intel and AMD cpus are MIA most places. The only real way to get them is to buy combos from newegg...which to be honest most people hate using newegg anymore.
Sure if you live in some big area. Meanwhile in the USA most people only have a walmart nearby. I'm not driving for 45min to the city and 45min back for something that may or may not be in stock.
you could always phone to see if they have it in stock, or, if possible order it online through them, and pick it up. the store i would go to, one location is about that for me as well, 45 mins, their other location, is about 30 mins
Amazon's been in stock for everything but the 3900 pretty much constantly for the last two weeks. And you can get the 3900 if you put it on notice and hop on as soon as you get the notice. They get shipments every so often but they sell out almost immediately on the 3900.
My bet is AMD decided to hold back on the release until they can saturate the market on the 3900 and stockpile enough chips to meet the initial demand. They've struggled to meet the retail demand because demands been so high and rather than disappoint people they probably decided it would be better to stockpile the chips for a month and a half so there was a solid initial supply on 3950 and TR3. I believe this is a reaction to the negative press they've had because they can't meet demand on the 3900 and initially even on the 3700 and 3800.
Consensus is that this is a supply issue...and there may be some of that. However, there is another possibility. They may just want Intel to show all its cards first either to move clocks slightly higher to insure they win the benchmarks or just to insure that their pricing makes them more attractive without giving too good of a bargain, so they can get some reasonable profit. I think there is reason to think this. They bragged that they got Nvidia to present their prices first. Just a couple months ago.
There is a 3rd possibility also. They could be continuing to refine BIOS stuff to insure they get the absolute best performance at release, when all the benchmarks are published. It is even a remote possibility that they are working with games companies to make patches that take better advantage of all those threads and the Zen 2 architecture, and they want those in place for the release to insure it overtakes Intel on games. And even if these things are not motivating this delay. The delay empowers them to do all these things. Good time to own AMD like me. I think they could easily be at $38 by the end of the year. 10,000 employees laying waste to 100,000. That is a story. And they are combating Nvidia too. They have yet to land a heavy blow...but I think that is inevitable. There was a bit of a coup recently where Windows directX 12 is going to handle ray-tracing regardless of CPU...if I understood that correctly.
The decision to purchase the 3900x was a no brainer. 50 percent more cores and double the cache for only $100 more then the 3800X. 3950x is a much harder sell at only 33 percent more cores and same amount of cache for $250 more. Threadripper might be interesting depending on the price points though.
No it doesn't. Total L2+L3 cache is 72MB (8MB of L2 and 64MB of L3). For comparison, the 3900X has 4 cores less, so at 0.5MB of L2/core it also sheds 2MB off the L2 for a total of 70MB.
Since Ryzen CPUs use L3 as a victim cache, it acts as an extension of L2, and therefore the numbers are sometimes added together. It's another matter entirely that each core only has access to the L3 attached to its CCX - 16MB. Each core's L2 caches will often contain some of the same data, as will different CCXes' L3 cache slices.
Also, since each CPU die has only 32MB of L3, there is no way you could get more with just two of them.
I think it's about time that AMD reinvented it's GHz rating. They should totally forget about their Boost clock and show the combined base-clock GHz.
The 3900X could be 45.6 combined GHz; The 3950X would be 56GHz.
Either way, you can't get done for misrepresentation here, and the best Intel could spin with it's 9900 series CPU's is 28GHz or 50GHz if it went with all-core turbo.
The point being, the numbers are all pretty meaningless anyway as there's significantly more to CPU power these days than just GHz.
beedoo, thats the sad part, if you notice, a lot of those that are praising intel, keep saying the same thing, intel can get 5ghz out of their chips, why cant AMD ?? or the intel chips clock faster so they give me an extra few frames in my games, that i probably dont really notice, so they are the better choice. its like the pentium 4 vs athlon/athlon XP days all over again.
Intel can only get 5 GHz out of it's very mature 14nm process. Smaller nodes give lower power consumption and not higher frequencies, at least in the early days. In fact 10nm and 7nm may never be able to clock as high as Intel's venerable 14nm. But, as was said above, there's more to CPU power than frequency - it's just that some people are obsessed with it. I actually quite like the GigaHertz-core proposal. Intel used to get irritated by the "plus" rating AMD used to give to its Athlon 64s and multiplying core count by base frequency would be a good response to Intel's recent presentation.
John_m yep.. but tell those that think the frequency is all that matters, that. and i think i was the A64 that was clocked lower then the P4 but still was faster, not the athlon/althon XP. i think an interesting comparison to do, would be to get to identical comps, less the mobo and cpu, one intel, one amd ( even if its just 1 cpu each, or multiple ) and clock them both the same, and set it up so the playing field is as level as possible, and see which one is faster at the same clocks...
One of the gaming-orientated YouTube channels (maybe Hardware Unboxed or Gamers' Nexus?) did that, locking them both to 4 GHz, and clearly demonstrated that AMD has the higher IPC.
Intel can get 5GHz because it's their own process. And they've spent most of the the last decade tuning it, along with their their transistor design, for maximum clocks. AMD can't really do that because they are using what is essentially an 'off the shelf' process, and a design that needs to be portable in case they need to swap foundries or make more somewhere else.
It's a trade off that AMD have had to make because they don't own their own process. So rather than worry about clocks so much, they've concentrated on IPC and core counts instead. Which is simply another way of increasing overall performance.
Much of that is process maturity, not ownership. After five years of TSMC 7nm, AMD would be able to hit 5.5GHz, but AMD isn't locked into the fab process for design improvements the way Intel did. There is a challenge in adding cores for the same power envelope, and Intel is still trying to figure out how to get beyond monolithic designs, because Intel has been complacent.
Kind of too bad that they're not leaving an upgrade path to Zen2 for those who bought into TR4 for the first gen 12-16 core parts. While I get that these parts wouldn't sell a lot, it kind of leaves people stranded unless they need heaps of cores. Of course, if the 24-core is >$1000 and boosts as high as the AM4 parts for lower threaded loads I guess that's not really an issues.
Why would you want to put an Epyc in a Threadripper motherboard? You'd throw away most of its advantages (8 channel memory and 128 PCIe lanes) and end up with a chip that costs a lot more and clocks slower. Put Epyc in a motherboard that's designed for it. In any case gen 3 Threadripper has been announced for launch in November.
I obviously meant no upgrade path without jumping to ridiculous core counts - TR4 compatibility for the upcoming Threadripper series is all but a given. 12-16 cores can be utilized for relatively common workstation tasks. 24 and above is another thing entirely. Personally I would hope for a high-boosting 16-core as the low end part of the next generation of TR - it would be a very sensible while still meaningful upgrade for any 1st-gen Threadripper system. But as I said, if the 24-core boosts as high as the Ryzen 7s and 9s, and uses the same boost mechanism (i.e. dynamically clocks higher when fewer cores are active) that could be a decent compromise.
I'll pass on both of them... Threadripper has never been a good platform for gaming, and the Ryzen 3950X is clearly having issues reaching it's promised boost clock... that's probably why it was delayed. Not to mention, the 3900X was introduced in July and it STILL isn't available at MSRP right now... It's a horrifically poor value.
9900K still is the gaming king and with hyper threading has well deserved it's i9 designation - just look at the Bench results compared to the 7900X. You can get the 9900k at microcenter.com for only $450 and then get another $30 off the motherboard... and most important ... it's ACTUALLY in stock!!
" Threadripper has never been a good platform for gaming " it never was meant to be a gaming platform. " Ryzen 3950X is clearly having issues reaching it's promised boost clock " and where did you see this ? post a source. " STILL isn't available at MSRP right now. " which could be due to demand, there has been other products that arent at msrp at release, and for a few months after, whats your point ? " It's a horrifically poor value. " where i am, its the same price as the 9900k how is that not a good value ? more cores, faster in multi threaded, and the few FPS you would get over the 9900k, most people, wont even notice, AND uses quite a bit LESS power " microcenter.com for only $450 " are you forgetting that you will NEED to get a cooler for this cpu ? and dont think that 50 buck cooler is going to cut it, you will need to get a cooler that is at least $70 or MORE to keep this cpu cool enough in order to get the performance you are raving about. the local comp stores here, usually have it in stock, so maybe stores there just cant get them in.
Not sure why you would want to compare the 3900x against the 9900k. The mere 3700x is pretty evenly matched against the 9900k, and is available for $349 with an appropriate cooler.
Price to price comparison. The 3900X is a $499 CPU(I don't count war profiteers increasing the prices to $750 for a $500 CPU). The i9-9900k is a $480 CPU that doesn't come with a cooler, so if anything, the 3900X is actually less expensive.
Multi-threaded edge goes to AMD with the 3900X, single threaded goes to Intel. Overall performance lead goes to the 3900X due to people doing more than a single thing with their computers.
not to mention, the single thread performance difference in games, most wont even be able to tell the difference. but the 9900x " should " be compared to its counterpart, the 3800x, which is also an 8 core, 16 thread cpu.
If Threadripper 3rd gen is 4 CCX, that's 32 cores(4x8). With AMD going to an I/O die, it is possible that AMD might go for more CCX in Threadripper, but I doubt it, because that would harm Epyc sales.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
82 Comments
Back to Article
Hifihedgehog - Friday, September 20, 2019 - link
https://www.youtube.com/watch?v=WWaLxFIVX1sGondalf - Friday, September 20, 2019 - link
Ummm hopefully this new top of the line SKU will be able to give the advetised single core turbo and all cores turbo and definitively not on only 5% of the stockpile of chips. Yes because if the crap story will repeat itself, nearly all the reviews on selected SKUs for press from AMD will have zero relevance, but it will be only a good marketing show. IMO the delay is for a more accurate binning at TSMC.Xyler94 - Friday, September 20, 2019 - link
First off, AMD never specified all-core turbos.Secondly, Intel mentions Max single core turbo also, but more as a guideline. Intel's chips never, under any circumstance, reach max turbo at their rated TDP. (Thermal or Power budget). Turbo speeds has always been "If power and cooling allow", even on Intel... especially on Intel too. And fun fact, usually to get the advertised turbo speeds on Intel, you can't be using any AVX, else you hit the offset.
There's only 1 guarantee, the base clocks. But who cares about a Ghz number, when what should matter is actual performance? AMD's FX9590 was 5ghz all core, but was still a crap CPU.
Jorgp2 - Sunday, September 22, 2019 - link
>Intel's chips never, under any circumstance, reach max turbo at their rated TDP.What's with AMD fanboys and spreading miainformation.
Silicon lottery is a thing, there's dies out there that can turbo at base TDP.
And most dual core Skylake U series CPUs will run at turbo indefinitely due to the fact their turbo clock is so low.
>Turbo speeds has always been "If power and cooling allow", even on Intel... especially on Intel too.
No, all Intel CPUs will turbo to some extent. The amount of time they spend in turbo is what changes.
Korguz - Sunday, September 22, 2019 - link
Jorgp2 too bad intels chips dont run at base TDP under use.i guess you havent read this : https://www.anandtech.com/show/13544/why-intel-pro... if you do limit intels cpus to their rated TDP, their performace, goes down the toliet. the 9900K for example, TDP of 95 watts, actual usage, 150-200 watts " No, all Intel CPUs will turbo to some extent. The amount of time they spend in turbo is what changes. " ONLY if the cpu has adequate cooling.Jorgp2 - Sunday, September 22, 2019 - link
You're either ignorant or misinformed.Korguz - Sunday, September 22, 2019 - link
if you didnt read the link, then you are ignorant AND misinformed. i take it you had NO idea how much power intels cpus REALLY use ?Netmsm - Monday, September 23, 2019 - link
I wonder how ignorant, but sure of misinformation one could be!olde94 - Tuesday, September 24, 2019 - link
i'll chip in and say that it's you. you will most likely not see any chips that can go full turbo at TDP in 2019. The difference between TDP and peak load are too far from each other and while the scilicon lottery is a thing, in most cases intel will let an amazin i5 upgrade to a low tier i7 and a low tier i7 upgrade to a high tier i7. they will push it further if it comes out as a great chip, so the only case where this would normally be seen would be on a top tier i7 that has just hit the golden everything, and it's too rare to talk about for most people.Xyler94 - Monday, September 23, 2019 - link
What's with Intel Fanboys instantly calling people AMD fanboys when they're simply stating things about Intel CPUs that are facts?And no, you're wrong on that front too. anything not a Core i5 or i7 doesn't have Turbo tech built in... unless I'm mistaken, I haven't purchased or seen a Core i3 since the Coffee Lake 8000 series...
Turbo speeds are never a guaranteed thing, Intel admits this. I'm no fanboy, I currently own no AMD devices, and I've got 4 computers to my name (Core i7 4790k and Xeon E3-1245-V5 make up the two most important PCs I own)
Targon - Tuesday, September 24, 2019 - link
You clearly don't understand the difference between what a chip can do, and what the TDP is. Going above base speeds will take more power, but how a chip is rated when it comes to TDP is very different between Intel and AMD.Intel ONLY reports the TDP for base speed. So if the i9-9900k says 95W TDP, that is at the base speed of 3.6GHz. The fact that the i9-9900ks has a TDP of 127W with a base speed of 4.0GHz indicates very clearly that just going from 3.6GHz to 4.0GHz boosts the TDP from 95W to 127W. So that magic 5.0GHz that people want isn't going to be at 95W, or 127W, it's going to be upwards of 200W.
AMD rates TDP for typical usage, and while that is a bit vague, it means that a 105W TDP CPU will be enough to run above base speeds and stay within the official TDP rating.
Again, Turbo/Boost speeds and the TDP rating are two different things. If power delivery to the CPU were limited to official numbers, you wouldn't see any Intel chips getting above 4.6-4.7GHz even with golden samples. Throw unlimited power at the chip, and you get higher clock speeds.
Sushisamurai - Saturday, September 21, 2019 - link
I think you should read the article that refers to how AMD's turbo's definition and how theirs works vs Intel, as its quite enlightening and I do agree with the observations made by Anandtech. I think AMD's done a great job and its nice to know that the CPU you're buying from them is pretty much at max performance, and could probably undervolt for better performance instead of overvolt and overclocking vs Intel's overclock for more performance because they're leaving more on the table for the user to exploit.Tabalan - Saturday, September 21, 2019 - link
Hmm, how did we get from "low OC headroom is a drawback" (Intel) or "all chips should be able to OC" (Intel) to "0 OC headroom is good" (AMD)?Also, OC is not an exploit, it's just a feature of almost all (AMD) or selected (Intel) SKUs. That's why you buy Intel K series - to Oc them.
Next thing - majority of AMD Ryzen 3000 CPUs don't hit Boost Clock ever. Watch De8auer video, hunders of people took part in survey and it turns out in some SKUs only few (<10) % of units hit Boost clock EVER.
teldar - Saturday, September 21, 2019 - link
I'm going to ask you to think about something. Who do you think is more likely to take part in a survey? Someone who is pissed off and has something to say? Or someone who is completely happy with what they have and have stopped thinking about it? It's a fact that voluntary surveys are skewed negative because of people are more like to complain if something isn't right than write about how good it is.I wouldn't take survey numbers like they're a scientific study.
Martijn ter Haar - Sunday, September 22, 2019 - link
There's that bias, but I also think Der Ba8er's survey has a problem, because of Windows Scheduler. AMD only guarantees the boost clock on one core, the one marked with the gold star in Ryzen Master. When I run the CB single core test on my Ryzen 3600 it does not get scheduled to that core, but to a seemingly random core. In fact, in spite of the name, I see loads peak on different threads during the run and I am pretty sure not all are on the same core (I use CB20 instead of CB15.) And there is no way to easily force Windows to use the gold star core (Maybe it can be done with Process Lasso?)If this is the case, it is no wonder that in Der Ba8er's survey the number of Ryzen 3000 CPU's that can't reach their advertised boost clock gets worse with a rising number of cores. The chance that the benchmark is run on the gold star core just gets lower.
Of course, this is not the entire story, because AMD's ABBA Agesa revision did help. But I still think Windows not assigning the CB single core benchmark run to the right, 'gold star' core plays an important role in Der Ba8er's survey results.
Korguz - Saturday, September 21, 2019 - link
tabalan... has de8auer ever talked about how much power intel uses to reach the clocks and performance it gets ? like teldar said.. those surveys are bias towards the negative. and should be taken as biasedFullmetalTitan - Saturday, September 21, 2019 - link
So it's clear from your reply that you DIDN'T read the article, because the impossibility of measuring an instantaneous boost without tainting the result is covered pretty thoroughly.WaltC - Saturday, September 21, 2019 - link
Next Thing is already here: with bioses built around AGESA ABBA, all Zen 2 CPUs are once again hitting their max advertised boosts MHz with ease--just as they did with with the early 1002 AGESAs--just like mine. How did we get to "overclocking is a guaranteed function of all cpus," ? when the *the fact* is just the opposite--overclocking is never guaranteed in any form by any CPU manufacturer. Oh--that includes Intel--which I guess I have to spell out. "OC headroom" is not something smart people care about--CPU performance, however, is very much something people care about. MHz, remember, is only half the performance picture--the other half--the more important part--is IPC execution. OC is not a "feature" guaranteed or warrantied" by any CPU manufacturer to any degree whatsoever. Sorry that you don't understand that OC is not a "feature"....*snicker*. You might be interested to know that 90% of the people who buy CPUs made by anyone--at least 90%-- do not bother OC'ing them at all.dgingeri - Saturday, September 21, 2019 - link
I think the biggest thing in this change is that we shouldn't have to overclock. With Intel, we've gotten used to them lowballing us and nerfing the chips to force us to overclock in order to get what we paid for, and people even made that into a sport. Overclocking became a 'feature' because Intel has been artificially sandbagging and cheating us for years.With the Zen chips, it is more of "you get what you paid for out of the box" kind of mentality and that different tiers of products actually offer valid upgrades for the extra money instead of just lesser degrees of artificial nerfing from the manufacturer.
Harry Voyager - Sunday, September 22, 2019 - link
I'm neutral on this. On the one hand, AMD chips not meeting their boost and having low head room is a drawback, however the performance we are seeing in the benchmarks is with those same lower boosts and lower headroom.Yes, it's a downer that they don't seem to hit the peak speeds, but at the same time, if you have a multi threaded load, they are still out performing Intel by a serious margins and nearly on par in single threaded loads. That is ultimately what matters.
WaltC - Saturday, September 21, 2019 - link
The demand for AMD's Zen2 has blown away all of the company's internal estimates--the demand for these fantastic CPUs exceeds AMD's studied projections. With demand skyrocketing, there is no way the original estimates could hold, because the nice thing about this from AMD's perspective is that their most expensive Zen2 CPUs are the most in demand! Seems to be a trend here.ValiumMm - Monday, September 23, 2019 - link
its already been fixed though, why you bringing this up?Spunjji - Monday, September 23, 2019 - link
Astroturfing or fanboy rage, never really clear which is which anymore. Both are annoying and harmful to anyone looking to actually learn something.eek2121 - Monday, September 23, 2019 - link
Ryzen chips are boosting just fne. That being said, I read somewhere the other day that the new Threadripper chips (or at least 1 of the SKUs) is using a new socket. I wonder if Ian has any insights into this?Smell This - Friday, September 20, 2019 - link
I guess this is a good/bad news kind-of-thang.
Launch is delayed, but demand is up (especially considering the enterprise side). Any input on demand of the various chiplets?
rahvin - Friday, September 20, 2019 - link
Glad to see them finally set a date. They've been saying Q4 for launch since the Q2 Financial call for TR3. I've been worried the retail side wouldn't see much because the cloud companies were sucking up all the silicon like the original Epyc that you couldn't buy for 6months after release because Google, Amazon and the others were soaking up the whole production run. I hope the same thing doesn't happen with TR3 and AMD has stockpiled enough to get it into the retail channel as well.You still can't really buy the 3900X on Amazon at the MSRP or below unless you get lucky and login right when a shipment arrives. I've only seen it available once at the MSRP. So AMD still has really strong demand for Ryzen Gen3 in the retail channel. I expect the 3950 will be even worse and the TR3 could suffer the same fate if they don't stockpile before release to retail.
catavalon21 - Saturday, September 21, 2019 - link
I agree. Having a product which is good enough for demand to strain supply is a good thing for AMD - especially if it's the very profitable EPYC line - just don't wait so long to fill retail channels that competition has time to catch up, or for consumer interest to wane.5080 - Friday, September 20, 2019 - link
This doesn't look good for future Ryzen APU's. Since there is already an issue with delivery times on TSMC's 7nm process for orders that have been placed months ago it will just get worse with all the mobile SoC (Qualcomm's 865 and Apple's A13) required over the next few months. It could take another year before we see any Ryzen 3 based APU's on the market.Flunk - Friday, September 20, 2019 - link
Ryzen 3 APUs are out now. They always fab the APU chips on the previous process with the previous core design. Ryzen 4 will be the first 7nm Zen 2 APUs.brucethemoose - Saturday, September 21, 2019 - link
The bulldozer APUs came out alongside the desktop CPUs with the same core generations, IIRC. Heck, the bulldozer cores got an extra iteration on mobile that the desktop never saw, as did K10 with the llano APUs.Jorgp2 - Sunday, September 22, 2019 - link
They did release carrizzo on FM2+, just not the fully unlocked versions.rocky12345 - Friday, September 20, 2019 - link
First thing I should say is that I am An AMD fan & every time they release a new gen of Ryzen always hope that they do well.No on to the part some won't like most likely. I do like the idea of the chiplet it makes tings more cost effective and just easier to make the CPU than going with a huge core setup like the other guys are doing. The issue I have is that when you have a CPU like the 3900x & 3950x that use multi chiplets. AMD needs to make sure that each chiplet is binned in such a way that at any point any core (single core Turbo) can reach the advertised speed for turbo boost.
From what I am seeing here is you may get one chiplet that has a few better cores in it but the other chiplet only has cores that will only get to just below the advertised turbo boost. They need to have binning that has both chiplets have cores in it that can do the max turbo. I am not saying oh they have to be able to full turbo on all cores at once what I am saying is that any core at any given time needs to be able to run at the max turbo speed not just a few select cores once and a while.
Heck even on my old i7 2600K all 4 cores can reach the max turbo with ease and those same 4 cores can reach up to 5.1GHz and this CPU is from what 2011-2012 time frame and made on 32nm process. No max turbo is not guaranteed just base clock is but when they advertise clock speeds and turbos those CPU's better be able to do them even if it is only for a few seconds at a time and every core should be able to do it not just a couple cores in the CPU now and then. A said I am talking about single core turbo not whole chip speed.
This is one of the reasons I am still on my Sandy bridge is the rather lack luster MHz on the Ryzen platform. Yes they perform well even at the lower clock rates. The other reason is that on the Intel side they priced themselves out of my wallet a long time ago oh and the ever changing socket game they like to play as well pretty much made me not want to have Intel in my system at least for now. This is just my opinion nothing more and nothing less.
guycoder - Friday, September 20, 2019 - link
I don't get your logic? AMD are only saying that turbo boost applies to a minimum of one core and even then no guarantee on how long that turbo will be for. As long as one core momentarily hits the advertised box speed then that is it. It doesn't matter which core or which chiplet so all AMD needs to make sure is one chiplet contains 1 core that can hit that turbo speed. I really think all these AMD/Intel fan boys have lost the plot and would be better off taking up another hobby such as drag racing. Pretty sad when humanity has come to the point where we can take a pile of sand though unimaginable technological and industrial processes just so some kid in his basement can run a stupid benchmark and then complain about a few missing mhz or a couple missing frames in a game..Mr.Vegas - Friday, September 20, 2019 - link
This is not charity, your comment sound like 'Wah Wah wah, dont say bad things about AMD'AMD is another greedy corp that wants to make money and we as Consumers that pay their bills should get the maximum, without if or maybes.
AMDs turbo is pure scam, they can easily write on the box 5Ghz and 6Ghz, as long as it hits the score for one nano second its OK.
The turbo should be useful not useless, the way it is now, AMDs turbo is advertising scam.
Korguz - Friday, September 20, 2019 - link
mr vegas. just like the power useage of intels cpus ?? thats a scam tooQasar - Friday, September 20, 2019 - link
mr vegas, i guess you havent seem and read this : https://www.anandtech.com/show/14873/reaching-for-...yeeeeman - Friday, September 20, 2019 - link
Amd specifies a single core turbo of 4.6Ghz for 3900x and that is reached. They never said that all core turbo is also 4.6Ghz. And nevertheless, it is not all about frequency. Even with this deficit 3700x and 9900k are even in compute tasksKorguz - Friday, September 20, 2019 - link
rocky12345 and the point is about your 2600k ?? so what if it can reach 5.1 ghz, almost any equivalent newer chip then that, would more then likely out perform it, even if clocked lower. what speed a cpu runs at, doesnt mean much, its the IPC the cpu does that is important, as for the comment on the turbo thing, did you read this : https://www.anandtech.com/show/14873/reaching-for-... ?????alufan - Saturday, September 21, 2019 - link
hmm an AMD fan who has a intel chip and bashes AMD? it has never been promised or stated that all cores could attain the max some do some dont, if you want real dishonesty look at the other team who claim TDPs that are frankly pure fantasy, I think we are also getting the result as stated on a few articles of the differing interpretation of the two big players when it comes to turbo Intel jacks up the power and gives all cores AMD stays within the TDP but only hits that max momentarily because the chip is more efficient to start with and simply works better at a given TDP if you restrict the chips to a given or stated TDP then see how well the Intel boosts it would paint a very different picture.FWIW the only CPU I have lost for no apparent reason was a team blue one a core 2 duo one, my one and only ever Intel purchase died after 2 years of modest use I was strictly AMD before or since, I just find them more honest and in daily use they have always felt snappier in a windows environment (which is probably purely in my head).
yeeeeman - Friday, September 20, 2019 - link
They probably didn't anticipate such a big success for Zen 2. They probably have the same situation on the server side with epyc so I am guessing they give priority to that which leaves consumer side with not much.yeeeeman - Friday, September 20, 2019 - link
Also I believe that amd gets a small chunk of tsmc production capacity since it is a new customer. It is a good problem to have high demand, but if you can't satisfy it then it means you lose money.rahvin - Monday, September 23, 2019 - link
I would wager TSMC prioritizes AMD wafers because they are a new client that has more wafer starts than any of their other clients.imaheadcase - Friday, September 20, 2019 - link
Rather have a small delay than no stock. Current high end Intel and AMD cpus are MIA most places. The only real way to get them is to buy combos from newegg...which to be honest most people hate using newegg anymore.Korguz - Friday, September 20, 2019 - link
heh.. i can go to one of my local comp stores right now, and get the ryzen cpus. keep in mind.. there are other stores besides amazon or new egg...imaheadcase - Friday, September 20, 2019 - link
Sure if you live in some big area. Meanwhile in the USA most people only have a walmart nearby. I'm not driving for 45min to the city and 45min back for something that may or may not be in stock.Korguz - Friday, September 20, 2019 - link
you could always phone to see if they have it in stock, or, if possible order it online through them, and pick it up. the store i would go to, one location is about that for me as well, 45 mins, their other location, is about 30 minsTrikkiedikkie - Thursday, September 26, 2019 - link
Do let people know where they can get the 3900Korguz - Thursday, September 26, 2019 - link
the store i go to, currently has 20 3900X @ $700 cdn each in stock according to their websiterahvin - Monday, September 23, 2019 - link
Amazon's been in stock for everything but the 3900 pretty much constantly for the last two weeks. And you can get the 3900 if you put it on notice and hop on as soon as you get the notice. They get shipments every so often but they sell out almost immediately on the 3900.My bet is AMD decided to hold back on the release until they can saturate the market on the 3900 and stockpile enough chips to meet the initial demand. They've struggled to meet the retail demand because demands been so high and rather than disappoint people they probably decided it would be better to stockpile the chips for a month and a half so there was a solid initial supply on 3950 and TR3. I believe this is a reaction to the negative press they've had because they can't meet demand on the 3900 and initially even on the 3700 and 3800.
Noodle-Naut - Friday, September 20, 2019 - link
Consensus is that this is a supply issue...and there may be some of that. However, there is another possibility. They may just want Intel to show all its cards first either to move clocks slightly higher to insure they win the benchmarks or just to insure that their pricing makes them more attractive without giving too good of a bargain, so they can get some reasonable profit. I think there is reason to think this. They bragged that they got Nvidia to present their prices first. Just a couple months ago.There is a 3rd possibility also. They could be continuing to refine BIOS stuff to insure they get the absolute best performance at release, when all the benchmarks are published. It is even a remote possibility that they are working with games companies to make patches that take better advantage of all those threads and the Zen 2 architecture, and they want those in place for the release to insure it overtakes Intel on games.
And even if these things are not motivating this delay. The delay empowers them to do all these things.
Good time to own AMD like me. I think they could easily be at $38 by the end of the year. 10,000 employees laying waste to 100,000. That is a story. And they are combating Nvidia too. They have yet to land a heavy blow...but I think that is inevitable. There was a bit of a coup recently where Windows directX 12 is going to handle ray-tracing regardless of CPU...if I understood that correctly.
Phynaz - Friday, September 20, 2019 - link
Another Month DelayAMD staying true to form.
catavalon21 - Sunday, September 22, 2019 - link
Another pointless comment. Phynaz staying true to form.Spunjji - Monday, September 23, 2019 - link
Now do Intel 10nmLiquidSilverZ - Monday, September 23, 2019 - link
It's Not True, Excuses, LiesPhynaz - Tuesday, September 24, 2019 - link
Good one!Trikkiedikkie - Thursday, September 26, 2019 - link
Similar to Intel then. With all their fabs and money they still cannot get the 10 going, raising prices to stupid levels and still ot able to deliver.The stupid ones are the people buying Intel
sygreenblum - Friday, September 20, 2019 - link
The decision to purchase the 3900x was a no brainer. 50 percent more cores and double the cache for only $100 more then the 3800X. 3950x is a much harder sell at only 33 percent more cores and same amount of cache for $250 more. Threadripper might be interesting depending on the price points though.JoeAceJR - Friday, September 20, 2019 - link
The 3950x has 72 megabytes of L3 cache. The 3900x has 64.Hul8 - Friday, September 20, 2019 - link
No it doesn't. Total L2+L3 cache is 72MB (8MB of L2 and 64MB of L3). For comparison, the 3900X has 4 cores less, so at 0.5MB of L2/core it also sheds 2MB off the L2 for a total of 70MB.AMD's site:
- 3900X: https://www.amd.com/en/product/8436
- 3950X: https://www.amd.com/en/products/cpu/amd-ryzen-9-39...
Since Ryzen CPUs use L3 as a victim cache, it acts as an extension of L2, and therefore the numbers are sometimes added together. It's another matter entirely that each core only has access to the L3 attached to its CCX - 16MB. Each core's L2 caches will often contain some of the same data, as will different CCXes' L3 cache slices.
Also, since each CPU die has only 32MB of L3, there is no way you could get more with just two of them.
beedoo - Saturday, September 21, 2019 - link
I think it's about time that AMD reinvented it's GHz rating. They should totally forget about their Boost clock and show the combined base-clock GHz.The 3900X could be 45.6 combined GHz; The 3950X would be 56GHz.
Either way, you can't get done for misrepresentation here, and the best Intel could spin with it's 9900 series CPU's is 28GHz or 50GHz if it went with all-core turbo.
The point being, the numbers are all pretty meaningless anyway as there's significantly more to CPU power these days than just GHz.
Korguz - Saturday, September 21, 2019 - link
beedoo, thats the sad part, if you notice, a lot of those that are praising intel, keep saying the same thing, intel can get 5ghz out of their chips, why cant AMD ?? or the intel chips clock faster so they give me an extra few frames in my games, that i probably dont really notice, so they are the better choice. its like the pentium 4 vs athlon/athlon XP days all over again.John_M - Saturday, September 21, 2019 - link
Intel can only get 5 GHz out of it's very mature 14nm process. Smaller nodes give lower power consumption and not higher frequencies, at least in the early days. In fact 10nm and 7nm may never be able to clock as high as Intel's venerable 14nm. But, as was said above, there's more to CPU power than frequency - it's just that some people are obsessed with it. I actually quite like the GigaHertz-core proposal. Intel used to get irritated by the "plus" rating AMD used to give to its Athlon 64s and multiplying core count by base frequency would be a good response to Intel's recent presentation.Korguz - Saturday, September 21, 2019 - link
John_m yep.. but tell those that think the frequency is all that matters, that. and i think i was the A64 that was clocked lower then the P4 but still was faster, not the athlon/althon XP. i think an interesting comparison to do, would be to get to identical comps, less the mobo and cpu, one intel, one amd ( even if its just 1 cpu each, or multiple ) and clock them both the same, and set it up so the playing field is as level as possible, and see which one is faster at the same clocks...John_M - Monday, September 23, 2019 - link
One of the gaming-orientated YouTube channels (maybe Hardware Unboxed or Gamers' Nexus?) did that, locking them both to 4 GHz, and clearly demonstrated that AMD has the higher IPC.John_M - Monday, September 23, 2019 - link
I had an Athlon 64 X2 6400+ (I've still got it actually, though it's retired now). It clocked two cores at 3200 MHz, hence the 6400+ designation.Haawser - Sunday, September 22, 2019 - link
Intel can get 5GHz because it's their own process. And they've spent most of the the last decade tuning it, along with their their transistor design, for maximum clocks. AMD can't really do that because they are using what is essentially an 'off the shelf' process, and a design that needs to be portable in case they need to swap foundries or make more somewhere else.It's a trade off that AMD have had to make because they don't own their own process. So rather than worry about clocks so much, they've concentrated on IPC and core counts instead. Which is simply another way of increasing overall performance.
Targon - Tuesday, September 24, 2019 - link
Much of that is process maturity, not ownership. After five years of TSMC 7nm, AMD would be able to hit 5.5GHz, but AMD isn't locked into the fab process for design improvements the way Intel did. There is a challenge in adding cores for the same power envelope, and Intel is still trying to figure out how to get beyond monolithic designs, because Intel has been complacent.Valantar - Sunday, September 22, 2019 - link
Kind of too bad that they're not leaving an upgrade path to Zen2 for those who bought into TR4 for the first gen 12-16 core parts. While I get that these parts wouldn't sell a lot, it kind of leaves people stranded unless they need heaps of cores. Of course, if the 24-core is >$1000 and boosts as high as the AM4 parts for lower threaded loads I guess that's not really an issues.Haawser - Sunday, September 22, 2019 - link
Where did anybody say that there wouldn't be an upgrade path ? Probably just need a bios update to make TR4 boards compatible with Rome.John_M - Monday, September 23, 2019 - link
Why would you want to put an Epyc in a Threadripper motherboard? You'd throw away most of its advantages (8 channel memory and 128 PCIe lanes) and end up with a chip that costs a lot more and clocks slower. Put Epyc in a motherboard that's designed for it. In any case gen 3 Threadripper has been announced for launch in November.Haawser - Monday, September 23, 2019 - link
My bad, I meant 7nm Threadripper, obviously.Valantar - Tuesday, September 24, 2019 - link
I obviously meant no upgrade path without jumping to ridiculous core counts - TR4 compatibility for the upcoming Threadripper series is all but a given. 12-16 cores can be utilized for relatively common workstation tasks. 24 and above is another thing entirely. Personally I would hope for a high-boosting 16-core as the low end part of the next generation of TR - it would be a very sensible while still meaningful upgrade for any 1st-gen Threadripper system. But as I said, if the 24-core boosts as high as the Ryzen 7s and 9s, and uses the same boost mechanism (i.e. dynamically clocks higher when fewer cores are active) that could be a decent compromise.TEAMSWITCHER - Sunday, September 22, 2019 - link
I'll pass on both of them... Threadripper has never been a good platform for gaming, and the Ryzen 3950X is clearly having issues reaching it's promised boost clock... that's probably why it was delayed. Not to mention, the 3900X was introduced in July and it STILL isn't available at MSRP right now... It's a horrifically poor value.9900K still is the gaming king and with hyper threading has well deserved it's i9 designation - just look at the Bench results compared to the 7900X. You can get the 9900k at microcenter.com for only $450 and then get another $30 off the motherboard... and most important ... it's ACTUALLY in stock!!
Korguz - Sunday, September 22, 2019 - link
" Threadripper has never been a good platform for gaming " it never was meant to be a gaming platform." Ryzen 3950X is clearly having issues reaching it's promised boost clock " and where did you see this ? post a source.
" STILL isn't available at MSRP right now. " which could be due to demand, there has been other products that arent at msrp at release, and for a few months after, whats your point ?
" It's a horrifically poor value. " where i am, its the same price as the 9900k how is that not a good value ? more cores, faster in multi threaded, and the few FPS you would get over the 9900k, most people, wont even notice, AND uses quite a bit LESS power
" microcenter.com for only $450 " are you forgetting that you will NEED to get a cooler for this cpu ? and dont think that 50 buck cooler is going to cut it, you will need to get a cooler that is at least $70 or MORE to keep this cpu cool enough in order to get the performance you are raving about. the local comp stores here, usually have it in stock, so maybe stores there just cant get them in.
catavalon21 - Sunday, September 22, 2019 - link
Not sure why you would want to compare the 3900x against the 9900k. The mere 3700x is pretty evenly matched against the 9900k, and is available for $349 with an appropriate cooler.Targon - Tuesday, September 24, 2019 - link
Price to price comparison. The 3900X is a $499 CPU(I don't count war profiteers increasing the prices to $750 for a $500 CPU). The i9-9900k is a $480 CPU that doesn't come with a cooler, so if anything, the 3900X is actually less expensive.Multi-threaded edge goes to AMD with the 3900X, single threaded goes to Intel. Overall performance lead goes to the 3900X due to people doing more than a single thing with their computers.
Trikkiedikkie - Thursday, September 26, 2019 - link
And single threaded is mostly games. Serious users want multi threadedKorguz - Thursday, September 26, 2019 - link
not to mention, the single thread performance difference in games, most wont even be able to tell the difference. but the 9900x " should " be compared to its counterpart, the 3800x, which is also an 8 core, 16 thread cpu.Spunjji - Monday, September 23, 2019 - link
Obvious disinformation shill comment is obvious.NICOXIS - Tuesday, September 24, 2019 - link
Will TR have a 64 core version?Targon - Tuesday, September 24, 2019 - link
If Threadripper 3rd gen is 4 CCX, that's 32 cores(4x8). With AMD going to an I/O die, it is possible that AMD might go for more CCX in Threadripper, but I doubt it, because that would harm Epyc sales.Trikkiedikkie - Thursday, September 26, 2019 - link
According to rumors one will be 64 core