Now, my curiosity is intrigued. Does this have a 4 lane PCIe gen4 link? Is this a harvested Tiger Lake U part, with the CPU mostly disabled? Is this dedicated silicon?
I still remember the i740.....So hyped, so good on paper, and it was the day we learned for GPU, or they were simply called Video Cards at the time, the quality of Drivers meant "Everything".
This time around Xe has all the Drivers work they have been working on with iGPU. Hopefully it will be a lot better. Although likely let down by their Fabs again.
Intel didn't have a choice but to make bad drivers. They totally lacked partner support with game developers which is why virtually no games ran properly on the i740. Synthetically it should have been competitive with the top cards of the time but there was some serious lack of communication between Intel and the community, which all considered, isn't really surprising. This is Intel we are talking about here...
I am really intrigued and would love to get my hands on this, I run a passive 1030 and both Nvidia and AMD don't cater to the low end space (Rumored 1010 coming, how exciting...).
This computer is used for photo work. If this can do better acceleration and performs comparable/better than a 1030, sign me up(Looking for ebay second hand I guess...). I seem to recall DG1 being around 1050 performance. Think of all the Ryzen workstations as well that could use this as a home as well thanks to the very much depleted and marked up GPU market we currently have (My GTX1080 sells for more than I paid for it 2 years ago). DG2 can't come soon enough!
There are no bad products, just bad prices. :) As it is OEM-only, perhaps we'll see a few on eBay or some Amazon sellers...
Your comparisons are a little off here, as this is a much lower power part, so it's best for SFF, HTPC, & OEM systems. 75W+ of passive heat would hamper thermals in most of these small cases.
1. 3x simultaneous display output 2. AV1 decode 3. Relevant & recent driver support 4. 2x to 4x the RAM (4GB vs 1GB / 2GB options) 5. GT1030 often exceeds TDP with 50W spikes (under 50ms) 6. Often ultra-cheap VRMs that throttle before the GPU
They reportedly require an Intel platform with specific BIOS support compatibility, so they may have effectively killed DG1 off for DIY users before it even hit retail. Sigh.
Oh, that's silly. Why? A PCIe dGPU is a PCIe dGPU the last I remembered. Does DG1 piggyback off the CPU iGPU in some hidden ways (thus it won't work with F-SKU CPUs)? Or Intel is very worried AMD HTPC owners might have another option?
Or is it PCIe incompatible, so they need vendor-specific shenanigans to avoid destroying their cards? Or its memory? Cheap GPUs have long used non-GDDRx VRAM. Or there are some Tiger Lake entrails left inside, so it'll muck up systems?
I still plan on buying on ebay and testing it unless I hear this more concretely. I expect there to be drivers out there to enable it jsut like they did for early DG1 samples they sent out for driver development.
Not silly at all. This is barely a working discrete part at all. Notice the LPDDR4X on a desktop card? That 4GB probably costs about as much as the same amount of GDDR6, if not more, but the memory controller is copy-pasted from TGL. The PCIe 4 lane limitation also sticks out. There's a simple answer here: it's not a PCIe dGPU. It's an onboard GPU decoupled from onboard fabric, still being initialized like it's an onboard GPU. It probably doesn't have a real vbios and needs to be initialized by the chipset.
It might not even work in every Intel machine, honestly.
The videocardz article suggests it only works on coffee lake S and comet lake S, only a selection of the mid range chipsets and only ones with a special bios. So it certainly looks like it doesn't work on many Intel machines.
But the performance delta of the DG1 over the 1030 is essentially zero.
And talking about SFF, you can use MSI Afterburner and GPU Boost on an GeForce to target a lower power threshold than the 75W TDP: https://sff.life/how-to-undervolt-gpu/
The tools are normally used for overclocking, but apply just as much to underclocking and undervolting. The article undervolts a 1070 -> 1060, shaving 54W TDP from a 1070.
I don't know if you can shave 35W from a 1650, but essentially if you limit the clock the GPU ever tries to hit, it won't ever reach the 75W TDP.
Given what Nvidia managed in transforming the 1650 into the MX450, I suspect you could happily chop 35W off a 1650 and get much better performance (and perf/watt) than the 1030. It would be a LOT slower than a stock 1650, though.
I was looking at upgrades for a nVidia 710 that was used for streaming. I noticed that the 1030/1010 doesn't support hardware accelerated video encoding. The 710 does. For that niche, the Intel DG1 or an AMD card might be the best. Unfortunately, the DG1 is OEM only.
Intriguing - Intel is willing to spend precious 10nm capacity on such low-margin parts. Or could it be 10nm isn't that tight anymore with AMD eating Intels market share slide-by-slice.
Vanilla 10nm - Ice Lake & Ice Lake SP 10nm SF - Tiger Lake U & H and this tiny GPU 10nm ESF - Golden Cove (Alder Lake, Sapphire Rapids and un named U SOC)
so other than Tiger Lake and maybe Xe HP that line has tons of capacity it seems
AMD is not eating anything and is hardly shipping anything for the PC market - TSMCs long fragile supply chain broken... Q1 is going to be a bloodbath for AMD...
"Q1 is going to be a bloodbath for AMD" Deicidium predictions are like Qanon drops: They never come true, and it never stops the true believers from buying in anyway.
These chips should be pretty small. And since they're selling them for laptops (where the perf/W advantage of 10 nm is most beneficial), it makes sense that the OEMs get the partially defective ones.
Give me one of these in half-height passive form and I'll stick it in my HTPC as soon as I can get my hands on it.
I'm having a hard time with my GK208-based GT730 not having certain video decode blocks (VP9 and AV1 are missing, h.265 is only partial) but I'm also having a hard time getting interested in a $100+ (CAD) 1030 that is essentially four years old at this point.
I wonder if Intel can produce a mining card with this GPU using 8 gigs of GDDR5. I don't know whether this is powerful enough to match the hashrates of an RX 480/570/580
It would, but not to anything that needs it. I think this thing flat out can't be initialized without the mobo chipset directly emulating how the CPU would initialize its onboard iGPU.
I would totally get a few of these for... my servers. Asus quality, Intel hardware, low power, passive cooling, 1 slot design. Hopefully a few of them wind up on eBay and I'll pick them up, either that or I hope Asus puts some of them out into the retail market.
Not to mention Linux device driver support, which tends to be rather good for Intel iGPUs...
...except perhaps for the newest variants including this.
I love the Iris Plus on my NUC home lab servers, because it works with every x86 OS on the planet and gives me a very reasonable KDE Plasma desktop at 4k. The NUC8i7BEH sold at €300 in October and November, which was a steal, and with 64GB of RAM and plenty of SSD storage they are generally very attractive.
Too bad Ryzen 4800U NUCs where never really available, I fear it will be the same with the 5800U while they cost (and likely deliver) double.
Frankly, I don't think you want that Vega 8 at 4K. It's not that great of a performer, and AMD's iGPU drivers on Linux aren't half as optimized as its dGPUs.
Supposedly this line supports GVT-g, Intel's proprietary alternative to SR-IOV. If true and supported by hypervisors this could have some interesting applications. Although it's a pretty anemic card, especially if you tried to split 4GB up between several VMs. Hopefully it's a sign of things to come with Intel releasing more impressive GPUs and lights a fire under AMD to release more mainstream mxGPU cards.
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
34 Comments
Back to Article
jeremyshaw - Tuesday, January 26, 2021 - link
Now, my curiosity is intrigued. Does this have a 4 lane PCIe gen4 link? Is this a harvested Tiger Lake U part, with the CPU mostly disabled? Is this dedicated silicon?Ryan Smith - Tuesday, January 26, 2021 - link
Yes, it's dedicated silicon (so no, it's not a harvested TGL-U part).And yes, it's just a 4 lane PCIe link.
Kurosaki - Tuesday, January 26, 2021 - link
Still avoiding the fun cards huh? ;)ksec - Tuesday, January 26, 2021 - link
I still remember the i740.....So hyped, so good on paper, and it was the day we learned for GPU, or they were simply called Video Cards at the time, the quality of Drivers meant "Everything".This time around Xe has all the Drivers work they have been working on with iGPU. Hopefully it will be a lot better. Although likely let down by their Fabs again.
Samus - Wednesday, January 27, 2021 - link
Intel didn't have a choice but to make bad drivers. They totally lacked partner support with game developers which is why virtually no games ran properly on the i740. Synthetically it should have been competitive with the top cards of the time but there was some serious lack of communication between Intel and the community, which all considered, isn't really surprising. This is Intel we are talking about here...cyrusfox - Tuesday, January 26, 2021 - link
I am really intrigued and would love to get my hands on this, I run a passive 1030 and both Nvidia and AMD don't cater to the low end space (Rumored 1010 coming, how exciting...).This computer is used for photo work. If this can do better acceleration and performs comparable/better than a 1030, sign me up(Looking for ebay second hand I guess...). I seem to recall DG1 being around 1050 performance. Think of all the Ryzen workstations as well that could use this as a home as well thanks to the very much depleted and marked up GPU market we currently have (My GTX1080 sells for more than I paid for it 2 years ago). DG2 can't come soon enough!
michael2k - Tuesday, January 26, 2021 - link
https://www.videocardbenchmark.net/compare/GeForce...The 1050 is almost 2x the Iris XE or 1030, and the Iris XE isn't really any faster than the 1030. The 1650 is 3x faster than the 1030.
There are passively cooled 1050 and 1650:
https://www.theverge.com/circuitbreaker/2017/2/6/1...
https://www.anandtech.com/show/15488/a-passivelyco...
ikjadoon - Tuesday, January 26, 2021 - link
There are no bad products, just bad prices. :) As it is OEM-only, perhaps we'll see a few on eBay or some Amazon sellers...Your comparisons are a little off here, as this is a much lower power part, so it's best for SFF, HTPC, & OEM systems. 75W+ of passive heat would hamper thermals in most of these small cases.
DG1 = 25W TDP
GT 1030 = 30W TDP (the actual comparison)
GTX 1050 = 75W TDP
GTX 1650 = 75W TDP
The benefits of DG1 over the GT 1030:
1. 3x simultaneous display output
2. AV1 decode
3. Relevant & recent driver support
4. 2x to 4x the RAM (4GB vs 1GB / 2GB options)
5. GT1030 often exceeds TDP with 50W spikes (under 50ms)
6. Often ultra-cheap VRMs that throttle before the GPU
A review: https://www.tomshardware.com/reviews/nvidia-geforc...
I don't expect *much* better from DG1, but I'd take DG1 over a GT1030 any day...
Slash3 - Tuesday, January 26, 2021 - link
They reportedly require an Intel platform with specific BIOS support compatibility, so they may have effectively killed DG1 off for DIY users before it even hit retail. Sigh.ikjadoon - Tuesday, January 26, 2021 - link
Oh, that's silly. Why? A PCIe dGPU is a PCIe dGPU the last I remembered. Does DG1 piggyback off the CPU iGPU in some hidden ways (thus it won't work with F-SKU CPUs)? Or Intel is very worried AMD HTPC owners might have another option?Or is it PCIe incompatible, so they need vendor-specific shenanigans to avoid destroying their cards? Or its memory? Cheap GPUs have long used non-GDDRx VRAM. Or there are some Tiger Lake entrails left inside, so it'll muck up systems?
What a silly restriction. Sigh x2.
cyrusfox - Wednesday, January 27, 2021 - link
I still plan on buying on ebay and testing it unless I hear this more concretely. I expect there to be drivers out there to enable it jsut like they did for early DG1 samples they sent out for driver development.lmcd - Thursday, January 28, 2021 - link
Not silly at all. This is barely a working discrete part at all. Notice the LPDDR4X on a desktop card? That 4GB probably costs about as much as the same amount of GDDR6, if not more, but the memory controller is copy-pasted from TGL. The PCIe 4 lane limitation also sticks out. There's a simple answer here: it's not a PCIe dGPU. It's an onboard GPU decoupled from onboard fabric, still being initialized like it's an onboard GPU. It probably doesn't have a real vbios and needs to be initialized by the chipset.It might not even work in every Intel machine, honestly.
PingSpike - Thursday, January 28, 2021 - link
The videocardz article suggests it only works on coffee lake S and comet lake S, only a selection of the mid range chipsets and only ones with a special bios. So it certainly looks like it doesn't work on many Intel machines.michael2k - Tuesday, January 26, 2021 - link
But the performance delta of the DG1 over the 1030 is essentially zero.And talking about SFF, you can use MSI Afterburner and GPU Boost on an GeForce to target a lower power threshold than the 75W TDP:
https://sff.life/how-to-undervolt-gpu/
The tools are normally used for overclocking, but apply just as much to underclocking and undervolting. The article undervolts a 1070 -> 1060, shaving 54W TDP from a 1070.
I don't know if you can shave 35W from a 1650, but essentially if you limit the clock the GPU ever tries to hit, it won't ever reach the 75W TDP.
Spunjji - Wednesday, February 3, 2021 - link
Given what Nvidia managed in transforming the 1650 into the MX450, I suspect you could happily chop 35W off a 1650 and get much better performance (and perf/watt) than the 1030. It would be a LOT slower than a stock 1650, though.jabber - Wednesday, January 27, 2021 - link
Buy a Quadro/FirePro on Ebay for very little.Rudde - Saturday, January 30, 2021 - link
I was looking at upgrades for a nVidia 710 that was used for streaming. I noticed that the 1030/1010 doesn't support hardware accelerated video encoding. The 710 does. For that niche, the Intel DG1 or an AMD card might be the best. Unfortunately, the DG1 is OEM only.ceisserer - Tuesday, January 26, 2021 - link
Intriguing - Intel is willing to spend precious 10nm capacity on such low-margin parts.Or could it be 10nm isn't that tight anymore with AMD eating Intels market share slide-by-slice.
Deicidium369 - Tuesday, January 26, 2021 - link
There are 3 DIFFERENT 10nm linesVanilla 10nm - Ice Lake & Ice Lake SP
10nm SF - Tiger Lake U & H and this tiny GPU
10nm ESF - Golden Cove (Alder Lake, Sapphire Rapids and un named U SOC)
so other than Tiger Lake and maybe Xe HP that line has tons of capacity it seems
AMD is not eating anything and is hardly shipping anything for the PC market - TSMCs long fragile supply chain broken... Q1 is going to be a bloodbath for AMD...
Spunjji - Wednesday, February 3, 2021 - link
"Q1 is going to be a bloodbath for AMD"Deicidium predictions are like Qanon drops: They never come true, and it never stops the true believers from buying in anyway.
zodiacfml - Wednesday, January 27, 2021 - link
because the integrated gpu was born from Tiger lake at 10nm. it would cost them to adapt this to 14nmKlimax - Wednesday, January 27, 2021 - link
And 14nm is oversubscribed as it is.mode_13h - Thursday, January 28, 2021 - link
These chips should be pretty small. And since they're selling them for laptops (where the perf/W advantage of 10 nm is most beneficial), it makes sense that the OEMs get the partially defective ones.powerarmour - Tuesday, January 26, 2021 - link
Should be decent for a HTPC or an older Linux workstation etc, purely for the video decode/encode capabilities (AV1 especially)evilspoons - Tuesday, January 26, 2021 - link
Give me one of these in half-height passive form and I'll stick it in my HTPC as soon as I can get my hands on it.I'm having a hard time with my GK208-based GT730 not having certain video decode blocks (VP9 and AV1 are missing, h.265 is only partial) but I'm also having a hard time getting interested in a $100+ (CAD) 1030 that is essentially four years old at this point.
zodiacfml - Wednesday, January 27, 2021 - link
I wonder if Intel can produce a mining card with this GPU using 8 gigs of GDDR5. I don't know whether this is powerful enough to match the hashrates of an RX 480/570/580Samus - Wednesday, January 27, 2021 - link
I'm curious if this card would add QuickSync support to a system.lmcd - Thursday, January 28, 2021 - link
It would, but not to anything that needs it. I think this thing flat out can't be initialized without the mobo chipset directly emulating how the CPU would initialize its onboard iGPU.bill.rookard - Wednesday, January 27, 2021 - link
I would totally get a few of these for... my servers. Asus quality, Intel hardware, low power, passive cooling, 1 slot design. Hopefully a few of them wind up on eBay and I'll pick them up, either that or I hope Asus puts some of them out into the retail market.abufrejoval - Wednesday, January 27, 2021 - link
Not to mention Linux device driver support, which tends to be rather good for Intel iGPUs......except perhaps for the newest variants including this.
I love the Iris Plus on my NUC home lab servers, because it works with every x86 OS on the planet and gives me a very reasonable KDE Plasma desktop at 4k. The NUC8i7BEH sold at €300 in October and November, which was a steal, and with 64GB of RAM and plenty of SSD storage they are generally very attractive.
Too bad Ryzen 4800U NUCs where never really available, I fear it will be the same with the 5800U while they cost (and likely deliver) double.
lmcd - Thursday, January 28, 2021 - link
Frankly, I don't think you want that Vega 8 at 4K. It's not that great of a performer, and AMD's iGPU drivers on Linux aren't half as optimized as its dGPUs.im.thatoneguy - Thursday, January 28, 2021 - link
Supposedly this line supports GVT-g, Intel's proprietary alternative to SR-IOV. If true and supported by hypervisors this could have some interesting applications. Although it's a pretty anemic card, especially if you tried to split 4GB up between several VMs. Hopefully it's a sign of things to come with Intel releasing more impressive GPUs and lights a fire under AMD to release more mainstream mxGPU cards.azfacea - Monday, February 1, 2021 - link
would I be surprised if both OEM and intel are lying and no such product has ever been shipped?? answer would be no I would notPaulHoule - Tuesday, February 2, 2021 - link
Never thought I'd see the day when "Intel Inside" became a warning label.