That would be a pricey HTPC processor. Although I guess you wouldn't really need a discrete card with this. I'm waiting for a review of the new Skull Canyon NUC.
I don't know, but Apple needs to hurry up and put out a Skylake refresh for more than just the Retina Macbook and also include a couple TB3 ports. I need a new laptop.
Maybe apple is waiting on Nvidia who is waiting on TSMC. Wouldn't that be funny if the reason Nvidia hasn't gotten access to 16 nm (Apple's SoC shipments) is the same reason they are late to refresh their macbooks? They're their own worst enemy.
Given current rumours about where the 2 known Polaris parts play, that's probably inaccurate. They should be able to cover all of the mid/high-mid and low-end with those.
Not if the iGPU is replacing the dGPU, but you would have to believe the power budget would be better spent with a lower power CPU and a dGPU unless you intend for the laptop to be paired with an external GPU when more GPU performance is needed, and the iGPU power budget is re-allocated to the CPU.
I think they would be, mobile quad core i7s are 47W in haswell, 45W in skylake. The Macbook Pro hasn't had a history of using desktop parts. And I'd agree; given laptops like the XPS 15, Apple is definitely falling behind in terms of GPU performance, and in actually updating Macbooks in general (aside from maybe the 12", which doesn't even have Thunderbolt 3 but I digress).
These are pure desktop parts,the rMBP would be using the H series alternatives to these. Like the i7 6770HQ. Same specs, just lower clocks all around. Really not bad chips for 45W.
The only thing with dGPU is the 5K iMac. What this means is that Apple could finally update the 21" to Skylake without adding a dGPU and keeping the 4K resolution.
Maybe theoretically possible in a current 15" MBP variation with no discrete graphics, but it incredibly unlikely, especially with a rumored redesign. These are likely to show up in new iMacs though.
45w is within the range of the larger MBP's TDP. I would think, even, that they could manage it in the smaller Pro as the "high-end" configuration -- other, smaller chassis managed higher combined CPU/discrete GPU TDPs, its just a matter of the cooling solution.
I'm hoping, personally, that this will be the case in the higher-spec'd 13" Pro, and that this will be the baseline (or minor upgrade) of the 15" Pro (where you get a discrete GPU in the high-end one). Historically, the 13" Pro has been a dual-core + HT machine, but frankly that's overdue to change. I'm also hoping they'll opt for the Mobile Xeon version of these chips, with ECC ram standard.
I believe these chips would also have configurable TDP down -- which I believe would be down to 35w in this case, and that's more than doable in a 13" chassis. It'd throttle itself, so you'd get less performance out of the 13", but that's still a lot of wallop in something so svelte.
This chip is the same config as the chip in the Skull Canyon NUC, and that'll play most contemporary games* at 1080p, 60FPS on medium/medium-high settings. The 72 EUs is 50% more than what's in the 3e configuration, and the eDRAM gives rather a lot of low-latency bandwidth.
*Those that don't routinely showcase as graphics benchmark showcases.
For some reason, no PC manufacturer seems to use this. The 15W TDP model is afaik used in one high-end version of XPS 13 and in expensive Surface 4 Pro variants. http://ark.intel.com/products/91163/Intel-Core-i7-...
If Apple is just doing a little internal update to their 13" model, the first one seems like a good candidate to be found in at least the BTO models. If they do more complete redesign, who knows what they'll pick.
Yeah those 28W ones are the 2+3e parts he was talking about. MUCH lower end chip, only 2 cores instead of 4, and only 48 instead of 72 EU's on the iGPU, and 64 instead of 128MB of eDRAM.
28W CPUs are used by apple in their rMBP 13. and a few others (asus, i believe)...
Everyone else uses 15W for everything... even most intel NUCs use 15W CPUs.
I hope intel makes a better version of the skull canyon NUC with this 65W CPU instead of the 45W CPU they used. We will accept a bit of height to fit a better cooler in there.
Purely at the level of the CPU these are too high as 35W - 47W is the usual range for laptop quads outside of the odd Xtreme edition or maybe workstation mobile CPU. If you look at the total TDP of a CPU + GPU then 65W is common although not sure what the actual max is for MacBook Pros.
Yes, but it's based on a 45W chip that is within range for the MBP. In the second picture, the red-highlighted version is actually the 45W "mobile" version - this new one is higher wattage.
This is a chip that is probably destined for the 21" iMac. (Hopefully the 27" retains the socketed CPU and discrete graphics.)
These are far, far outside the MBP's CPU's TDP range. Apple would never consider them for a laptop. They'd be a great choice for an iMac refresh, potentially, but then, since Apple typically puts "full-on" desktop-class CPUs in those units, something like this would be a step down if anything. The good news is, there are indeed 45W-class CPUs in Skylake that have Iris Pro GPUs, something Apple will certainly want to put into their entry level 15" models.
It really is. It's frustrating. It's like when you want to support someone but they'd always jumps on your shoulders on the slightest smile you give them...
Intel officially calls it eDRAM, whether it's a multi-die package or not. It's the modular nature of this feature, rather than create another two separate monolithic dies in their product stack. Even if you consider this to not be 'a true eDRAM' representation, it acts as one for all intents and purposes. If it was integrated on die, chances are it would be made into a true L4, but it would also have to be designed into a combined rectangular die for production efficiency - with this modular outset, it's just the internal on-package interconnects that have to be arranged properly.
Ian is the real deal! You have been and continue to be one of my favorite journalists at this site. You don't just write about it, you actually do it, and I love that you are willing to respond to readers to provide clarification etc. Keep up the good work Ian!
For a while we were creating the EDRAM on the same exposure field as the processors. The first four dies in the reticle array were the CPU and the 5th was the EDRAM. We stopped doing that because the yield impact was just far too high and the parts were physically separated in sort anyways.
L3 was not "disabled" (on the 6MB L3 version at least) on GT3e Haswell/Broadwell. Rather, a portion of it was used as the tag bits for the L4. This was mentioned by Anandtech before, it can be clearly seen in the diagram as well - http://www.anandtech.com/show/9582/intel-skylake-m... So 8MB L3 being available now is a result of the L4 no longer being directly attached to the L3 as a victim cache - it now got its own tag bits.
I know that image for sure (I wrote the page you linked to), was never aware that the 2MB of cache being 'eaten' was purely for the tags and had assumed it was just a decision made based on the die. It makes sense, and the image can certainly be interpreted that way.
You can see from Surface Pro 4's review that it takes that long to reach full frequency. It doesn't go from idle to full right away, but rather in steps.
You're right. I have this mental jump every time I write about Speed Shift. 'Shouldn't it be microseconds?' and I remember a discussion with Intel about this as well.
I'd be interested in one, too (well, the i7), but it must not appear shortly before the 10 nm update. As it was with Broadwell & Skylake.
And I have to be able to put that iGPU to good use. On my Ivy Bridge HD4000 I was happily crunching Einstein@Home, but on my current HD530 that doesn't work, apparently due to Intel driver bugs (no crashes, just completely wrong results). before this is fixed a beefier iGPU wouldn't make sense for me.
About frikkin' time Intel! If this means I can finally get Radeon HD5770 level of performance without resorting to discrete GPU solution and cool my entire computer using a single high-quality cooler I'm definitely grabbing one (i5-6685R that is). No more annoying GPU fans ...
Sorry AMD, your Raven Ridge (Zen-based) is taking too long.
i still cant buy a broadwell cpu let alone one with Edram why cant they make 2 high end I5 and i7 with the Edram and let us have a cpu that really does give more than 5%. that is the only cpu in broadwell and skylake that give you tru next gen and about 12-15% improvement
The i7-6785R is the processor I would have loved to see in Intel's latest Skull Canyon NUC (with the understanding that a different cooling arrangement might have been required). It is even cheaper than the one used there. This processor could be the base for a small competent system. I hope somebody picks it up and builds a nice compact system around it.
Ian, would it be possible to do an investigative benchmark into "Speed Shift" on windows 10, and the support for HWP in the Linux kernel? It seems they are the same, but I would like to understand if Windows 10 has a real advantage over the Linux kernel support. According to https://events.linuxfoundation.org/sites/events/fi... on page 25, these two things are the same (and the PDF above is from Intel). HWP in the intel_pstate cpufreq driver has been around since 2014, and has continued to be updated since. In fact, Linux 4.6 will have it enabled by default if hardware supports it.
Hi, first post so hopefully not to far OT with TDP & some of the portable Intel processor discussion.
Intel Portable i7 6650U @ 3.4 GHZ with 2 cores & 540 GPU TPD is rated at 15W , i7 6700HQ @3.5Ghz 4 cores & 530 GPU is rated at 45W TDP. With all the other SOC components which would not scale 2 -> 4 with the cores why does adding 2 more cores triple instead of double the TDP?
If it the density of the components can you correlate what expected power usage would be?
That wattage is misleading however. Keep in mind the i7 6700hq can keep a high CPU clock speed while the gpu is at full load/speed, while the 6650u will throttle to about 1.2-1.4 GHz and the gpu will throttle to 500~MHz when both are at full load.
Thanks explains a lot. Didn’t realize the Ghz TDP was peak and not sustained as I thought short term peak was “turbo”. Wonder if the M7 6Y75 2.9Ghz @ 7 watt TDP suffers similar throttling as it is normally not cooled beyond thermal attachment to a metal case.
A newer portable processor I’ve been interested in the i7-6567U Base 3.3GHz TDP @ 28 watts. I wonder if this indicates the processor can function sustained (2x?) higher Ghz than the 15Watt 6650U when properly cooled.
Obtaining Psychology Paper Writing Services online is fulfilling for most students who have to struggle with deadlines on Psychology Essay Writing Services and Psychology Research Paper Writing Services. https://www.meldaresearch.com/psychology-writing-s...
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
67 Comments
Back to Article
Eden-K121D - Monday, May 2, 2016 - link
Hmmm HTPC's would be a good use caseragenalien - Monday, May 2, 2016 - link
That would be a pricey HTPC processor. Although I guess you wouldn't really need a discrete card with this. I'm waiting for a review of the new Skull Canyon NUC.Eden-K121D - Tuesday, May 3, 2016 - link
Yeah That One looks promising.If they could somehow fit this into skull canyon it would be sweetDeath666Angel - Tuesday, May 3, 2016 - link
At this point, I would not buy an expensive HTPC, because of the lack of HDMI 2.0 options on the discreet and integrated GPU market.Murloc - Tuesday, May 3, 2016 - link
I've read that's coming with kaby lakevFunct - Monday, May 2, 2016 - link
Are these outside of the range of MacBook Pros TDP?cygnus1 - Monday, May 2, 2016 - link
I don't know, but Apple needs to hurry up and put out a Skylake refresh for more than just the Retina Macbook and also include a couple TB3 ports. I need a new laptop.willis936 - Monday, May 2, 2016 - link
Maybe apple is waiting on Nvidia who is waiting on TSMC. Wouldn't that be funny if the reason Nvidia hasn't gotten access to 16 nm (Apple's SoC shipments) is the same reason they are late to refresh their macbooks? They're their own worst enemy.jasonelmore - Monday, May 2, 2016 - link
This, Apple is not gonna release another 15" MBP with the same GPU.. They have effectively been on the same GPU since the 650m (performance wise)ragenalien - Monday, May 2, 2016 - link
The current rumor mill has polaris in the next macbooks.Eden-K121D - Tuesday, May 3, 2016 - link
I heard that most of R9 400m series would be rebranded except for maybe 2-3 high end partsSpunjji - Tuesday, May 3, 2016 - link
Given current rumours about where the 2 known Polaris parts play, that's probably inaccurate. They should be able to cover all of the mid/high-mid and low-end with those.Rampart19 - Monday, May 2, 2016 - link
These would be for the iMacs.onewingedangel - Monday, May 2, 2016 - link
Not if the iGPU is replacing the dGPU, but you would have to believe the power budget would be better spent with a lower power CPU and a dGPU unless you intend for the laptop to be paired with an external GPU when more GPU performance is needed, and the iGPU power budget is re-allocated to the CPU.rev3rsor - Monday, May 2, 2016 - link
I think they would be, mobile quad core i7s are 47W in haswell, 45W in skylake. The Macbook Pro hasn't had a history of using desktop parts. And I'd agree; given laptops like the XPS 15, Apple is definitely falling behind in terms of GPU performance, and in actually updating Macbooks in general (aside from maybe the 12", which doesn't even have Thunderbolt 3 but I digress).Valantar - Monday, May 2, 2016 - link
These are pure desktop parts,the rMBP would be using the H series alternatives to these. Like the i7 6770HQ. Same specs, just lower clocks all around. Really not bad chips for 45W.extide - Tuesday, May 3, 2016 - link
6770HQ is NO Iris Pro/Crystalwellextide - Tuesday, May 3, 2016 - link
NOT*dezonio2 - Tuesday, May 3, 2016 - link
6770HQ has the Iris Pro 580 with 128 eDRAM.http://ark.intel.com/products/93341/Intel-Core-i7-...
wallysb01 - Tuesday, May 3, 2016 - link
The only thing with dGPU is the 5K iMac. What this means is that Apple could finally update the 21" to Skylake without adding a dGPU and keeping the 4K resolution.freeskier93 - Monday, May 2, 2016 - link
Maybe theoretically possible in a current 15" MBP variation with no discrete graphics, but it incredibly unlikely, especially with a rumored redesign. These are likely to show up in new iMacs though.ravyne - Monday, May 2, 2016 - link
45w is within the range of the larger MBP's TDP. I would think, even, that they could manage it in the smaller Pro as the "high-end" configuration -- other, smaller chassis managed higher combined CPU/discrete GPU TDPs, its just a matter of the cooling solution.I'm hoping, personally, that this will be the case in the higher-spec'd 13" Pro, and that this will be the baseline (or minor upgrade) of the 15" Pro (where you get a discrete GPU in the high-end one). Historically, the 13" Pro has been a dual-core + HT machine, but frankly that's overdue to change. I'm also hoping they'll opt for the Mobile Xeon version of these chips, with ECC ram standard.
ravyne - Monday, May 2, 2016 - link
I believe these chips would also have configurable TDP down -- which I believe would be down to 35w in this case, and that's more than doable in a 13" chassis. It'd throttle itself, so you'd get less performance out of the 13", but that's still a lot of wallop in something so svelte.This chip is the same config as the chip in the Skull Canyon NUC, and that'll play most contemporary games* at 1080p, 60FPS on medium/medium-high settings. The 72 EUs is 50% more than what's in the 3e configuration, and the eDRAM gives rather a lot of low-latency bandwidth.
*Those that don't routinely showcase as graphics benchmark showcases.
zepi - Tuesday, May 3, 2016 - link
For 13" laptops Intel offers a 28W TDP model with slightly reduced edram:http://ark.intel.com/products/91167/Intel-Core-i7-...
For some reason, no PC manufacturer seems to use this. The 15W TDP model is afaik used in one high-end version of XPS 13 and in expensive Surface 4 Pro variants.
http://ark.intel.com/products/91163/Intel-Core-i7-...
If Apple is just doing a little internal update to their 13" model, the first one seems like a good candidate to be found in at least the BTO models. If they do more complete redesign, who knows what they'll pick.
extide - Tuesday, May 3, 2016 - link
Yeah those 28W ones are the 2+3e parts he was talking about. MUCH lower end chip, only 2 cores instead of 4, and only 48 instead of 72 EU's on the iGPU, and 64 instead of 128MB of eDRAM.8steve8 - Wednesday, May 4, 2016 - link
28W CPUs are used by apple in their rMBP 13.and a few others (asus, i believe)...
Everyone else uses 15W for everything... even most intel NUCs use 15W CPUs.
I hope intel makes a better version of the skull canyon NUC with this 65W CPU instead of the 45W CPU they used. We will accept a bit of height to fit a better cooler in there.
smilingcrow - Tuesday, May 3, 2016 - link
Purely at the level of the CPU these are too high as 35W - 47W is the usual range for laptop quads outside of the odd Xtreme edition or maybe workstation mobile CPU.If you look at the total TDP of a CPU + GPU then 65W is common although not sure what the actual max is for MacBook Pros.
CharonPDX - Thursday, May 12, 2016 - link
Yes, but it's based on a 45W chip that is within range for the MBP. In the second picture, the red-highlighted version is actually the 45W "mobile" version - this new one is higher wattage.This is a chip that is probably destined for the 21" iMac. (Hopefully the 27" retains the socketed CPU and discrete graphics.)
cmasontaylor - Thursday, May 19, 2016 - link
These are far, far outside the MBP's CPU's TDP range. Apple would never consider them for a laptop. They'd be a great choice for an iMac refresh, potentially, but then, since Apple typically puts "full-on" desktop-class CPUs in those units, something like this would be a step down if anything. The good news is, there are indeed 45W-class CPUs in Skylake that have Iris Pro GPUs, something Apple will certainly want to put into their entry level 15" models.madwolfa - Monday, May 2, 2016 - link
Come on, AnandTech. Full screen pop up ads? I'd have to reconsider turning off my AdBlock here.ddriver - Monday, May 2, 2016 - link
AT without addblock is just too disgusting, with addblock only moderately so :)madwolfa - Monday, May 2, 2016 - link
I'm trying to support them, but this was over the top.lilmoe - Tuesday, May 3, 2016 - link
It really is. It's frustrating. It's like when you want to support someone but they'd always jumps on your shoulders on the slightest smile you give them...wow&wow - Monday, May 2, 2016 - link
"Intel adds ... eDRAM"eDRAM is for DRAM that is integrated on the same die, i.e. monolithic.
Intel's solution is call hybrid, multiple dies on the same package substrate.
Others may be able to differentiate, but Anatech should : (
wow&wow - Monday, May 2, 2016 - link
Oops, correction:Others may not be able to differentiate, but Anatech should : (
Ian Cutress - Tuesday, May 3, 2016 - link
Intel officially calls it eDRAM, whether it's a multi-die package or not.It's the modular nature of this feature, rather than create another two separate monolithic dies in their product stack. Even if you consider this to not be 'a true eDRAM' representation, it acts as one for all intents and purposes. If it was integrated on die, chances are it would be made into a true L4, but it would also have to be designed into a combined rectangular die for production efficiency - with this modular outset, it's just the internal on-package interconnects that have to be arranged properly.
Also, Anandtech.
fanofanand - Tuesday, May 3, 2016 - link
Ian is the real deal! You have been and continue to be one of my favorite journalists at this site. You don't just write about it, you actually do it, and I love that you are willing to respond to readers to provide clarification etc. Keep up the good work Ian!JKflipflop98 - Wednesday, May 11, 2016 - link
For a while we were creating the EDRAM on the same exposure field as the processors. The first four dies in the reticle array were the CPU and the 5th was the EDRAM. We stopped doing that because the yield impact was just far too high and the parts were physically separated in sort anyways.mczak - Monday, May 2, 2016 - link
L3 was not "disabled" (on the 6MB L3 version at least) on GT3e Haswell/Broadwell. Rather, a portion of it was used as the tag bits for the L4. This was mentioned by Anandtech before, it can be clearly seen in the diagram as well - http://www.anandtech.com/show/9582/intel-skylake-m... So 8MB L3 being available now is a result of the L4 no longer being directly attached to the L3 as a victim cache - it now got its own tag bits.mczak - Monday, May 2, 2016 - link
Hmm, stupid logic detecting links: http://www.anandtech.com/show/9582/intel-skylake-m...Ian Cutress - Tuesday, May 3, 2016 - link
I know that image for sure (I wrote the page you linked to), was never aware that the 2MB of cache being 'eaten' was purely for the tags and had assumed it was just a decision made based on the die. It makes sense, and the image can certainly be interpreted that way.tuxRoller - Monday, May 2, 2016 - link
Correction:Milliseconds not microseconds with regards to Speed Shift.
djayjp - Monday, May 2, 2016 - link
So processors prior to skylake took up to 100ms? A tenth of a second huh. Certainly not, it's microseconds.IntelUser2000 - Tuesday, May 3, 2016 - link
You wouldn't notice microsecond differences so much.Yes, it is milliseconds: http://images.anandtech.com/doci/9751/SpeedShift.p...
You can see from Surface Pro 4's review that it takes that long to reach full frequency. It doesn't go from idle to full right away, but rather in steps.
Ian Cutress - Tuesday, May 3, 2016 - link
You're right. I have this mental jump every time I write about Speed Shift. 'Shouldn't it be microseconds?' and I remember a discussion with Intel about this as well.djayjp - Monday, May 2, 2016 - link
"50Gbps"... surely you mean 50GBps (giving a total of up to 100GB/s of bandwidth to the eDRAM).Ian Cutress - Tuesday, May 3, 2016 - link
You are right, corrected!beginner99 - Tuesday, May 3, 2016 - link
When do we get the 5775c successor?Ian Cutress - Tuesday, May 3, 2016 - link
I'd love to see one. Potential buyers need to make noise to Intel that plenty of people are willing to buy it though.MrSpadge - Tuesday, May 3, 2016 - link
I'd be interested in one, too (well, the i7), but it must not appear shortly before the 10 nm update. As it was with Broadwell & Skylake.And I have to be able to put that iGPU to good use. On my Ivy Bridge HD4000 I was happily crunching Einstein@Home, but on my current HD530 that doesn't work, apparently due to Intel driver bugs (no crashes, just completely wrong results). before this is fixed a beefier iGPU wouldn't make sense for me.
extide - Tuesday, May 3, 2016 - link
Well... the 5775c came out right AFTER a new process, not right before (and it uses the new process..) It came out right before a new arch, though.Arnulf - Tuesday, May 3, 2016 - link
About frikkin' time Intel! If this means I can finally get Radeon HD5770 level of performance without resorting to discrete GPU solution and cool my entire computer using a single high-quality cooler I'm definitely grabbing one (i5-6685R that is). No more annoying GPU fans ...Sorry AMD, your Raven Ridge (Zen-based) is taking too long.
nunya112 - Tuesday, May 3, 2016 - link
i still cant buy a broadwell cpu let alone one with Edram why cant they make 2 high end I5 and i7 with the Edram and let us have a cpu that really does give more than 5%. that is the only cpu in broadwell and skylake that give you tru next gen and about 12-15% improvementMobile-Dom - Tuesday, May 3, 2016 - link
It's a shame these arent LGA and socketed, because these would (arely) fit inside the thermal limit of the new MiniSTX form factor.a Hackintosh mini based on other of those would be nuts.
nomad73 - Tuesday, May 3, 2016 - link
The i7-6785R is the processor I would have loved to see in Intel's latest Skull Canyon NUC (with the understanding that a different cooling arrangement might have been required). It is even cheaper than the one used there. This processor could be the base for a small competent system. I hope somebody picks it up and builds a nice compact system around it.Sushisamurai - Wednesday, May 4, 2016 - link
When are we going to get a review? Are these over clockable (because that would be awesome)?Sushisamurai - Wednesday, May 4, 2016 - link
so these are not socketed/LGA are they... ... :(Meaker10 - Wednesday, May 4, 2016 - link
Correct me if i'm wrong but much like the 4xxx series the big bit is the cache and the small bit is the CPU ;)Drazick - Wednesday, May 4, 2016 - link
I wish Intel created High TDP socket CPU with the full Power of Iris Pro.Namely something like 6770K with this GPU.
I'd even go farther, give us Skylake-E with this built in GPU and eDRAM.
This will simplify a lot building a great computational system.
Great workstation for MATLAB, Julya, Python, etc...
alpha64 - Wednesday, May 4, 2016 - link
Ian, would it be possible to do an investigative benchmark into "Speed Shift" on windows 10, and the support for HWP in the Linux kernel? It seems they are the same, but I would like to understand if Windows 10 has a real advantage over the Linux kernel support. According to https://events.linuxfoundation.org/sites/events/fi... on page 25, these two things are the same (and the PDF above is from Intel). HWP in the intel_pstate cpufreq driver has been around since 2014, and has continued to be updated since. In fact, Linux 4.6 will have it enabled by default if hardware supports it.alpha64 - Wednesday, May 4, 2016 - link
Sorry, meant Anton...skidaddy - Thursday, May 5, 2016 - link
Hi, first post so hopefully not to far OT with TDP & some of the portable Intel processor discussion.Intel Portable i7 6650U @ 3.4 GHZ with 2 cores & 540 GPU TPD is rated at 15W , i7 6700HQ @3.5Ghz 4 cores & 530 GPU is rated at 45W TDP. With all the other SOC components which would not scale 2 -> 4 with the cores why does adding 2 more cores triple instead of double the TDP?
If it the density of the components can you correlate what expected power usage would be?
thanks
TheinsanegamerN - Wednesday, May 11, 2016 - link
That wattage is misleading however. Keep in mind the i7 6700hq can keep a high CPU clock speed while the gpu is at full load/speed, while the 6650u will throttle to about 1.2-1.4 GHz and the gpu will throttle to 500~MHz when both are at full load.skidaddy - Wednesday, May 25, 2016 - link
Thanks explains a lot. Didn’t realize the Ghz TDP was peak and not sustained as I thought short term peak was “turbo”. Wonder if the M7 6Y75 2.9Ghz @ 7 watt TDP suffers similar throttling as it is normally not cooled beyond thermal attachment to a metal case.A newer portable processor I’ve been interested in the i7-6567U Base 3.3GHz TDP @ 28 watts. I wonder if this indicates the processor can function sustained (2x?) higher Ghz than the 15Watt 6650U when properly cooled.
Oxford Guy - Thursday, May 12, 2016 - link
Kaby Lake.K-amps - Sunday, May 29, 2016 - link
Has intel given an indication of when they plan to release extreme Skylake parts to replace the 58xx or 59xx parts? i.e. Hex or oct i7 cpusdorcassmith - Monday, March 2, 2020 - link
Obtaining Psychology Paper Writing Services online is fulfilling for most students who have to struggle with deadlines on Psychology Essay Writing Services and Psychology Research Paper Writing Services.https://www.meldaresearch.com/psychology-writing-s...