The AMD Radeon VII Review: An Unexpected Shot At The High-End
by Nate Oh on February 7, 2019 9:00 AM ESTFP64 Performance and Separating Radeon VII from Radeon Instinct MI50
One of the interesting and amusing consequences of the Radeon VII launch is that for the first time in quite a while, AMD has needed to seriously think about how they’re going to differentiate their consumer products from their workstation/server products. While AMD has continued to offer workstation and server hardware via the Radeon Pro and Radeon Instinct series, the Vega 20 GPU is AMD’s first real server-grade GPU in far too long. So, while those products were largely differentiated by the software features added to their underlying consumer-grade GPUs, Radeon VII brings some new features that aren’t strictly necessary for consumers.
It may sound like a trivial matter – clearly AMD should just leave everything enabled – but as the company is trying to push into the higher margin server business, prosumer products like the Radeon VII are in fact a tricky proposition. AMD needs to lock away enough of the server functionality of the Vega 20 GPU that they aren’t selling the equivalent of a Radeon Instinct MI50 for a fraction of the price. On the other hand, it’s in their interest to expose some of these features in order to make the Radeon VII a valuable card in its own right (one that can justify a $699 price tag), and to give developers a taste of what AMD’s server hardware can do.
Case in point is the matter of FP64 performance. As we noted in our look at the Vega 20 GPU, Vega 20’s FP64 performance is very fast: it’s one-half the FP32 rate, or 6.9 TFLOPS. This is one of the premium features of Vega 20, and since Radeon VII was first announced back at CES, the company has been struggling a bit to decide how much of that performance to actually make available to the Radeon VII. At the time of its announcement, we were told that the Radeon VII would have unrestricted (1/2) FP64 performance, only to later be told that it would be 1/8. Now, with the actual launch of the card upon us, AMD has made their decision: they’ve split it down the middle and are doing a 1/4 rate.
Looking to clear things up, AMD put out a statement:
The Radeon VII graphics card was created for gamers and creators, enthusiasts and early adopters. Given the broader market Radeon VII is targeting, we were considering different levels of FP64 performance. We previously communicated that Radeon VII provides 0.88 TFLOPS (DP=1/16 SP). However based on customer interest and feedback we wanted to let you know that we have decided to increase double precision compute performance to 3.52 3.46 TFLOPS (DP=1/4SP).
If you looked at FP64 performance in your testing, you may have seen this performance increase as the VBIOS and press drivers we shared with reviewers were pre-release test drivers that had these values already set. In addition, we have updated other numbers to reflect the achievable peak frequency in calculating Radeon VII performance as noted in the [charts].
The end result is that while the Radeon VII won’t be as fast as the MI60/MI50 when it comes to FP64 compute, AMD is going to offer the next best thing, just one step down from those cards.
At 3.5 TLFLOPS of theoretical FP64 performance, the Radeon VII is in a league of its own for the price. There simply aren’t any other current-generation cards priced below $2000 that even attempt to address the matter. All of NVIDIA’s GeForce cards and all of AMD’s other Radeon cards straight-up lack the necessary hardware for fast FP64. The next closest competitor to the Radeon VII in this regard is NVIDIA’s Titan V, at more than 4x the price.
It’s admittedly a bit of a niche market, especially when so much of the broader industry focus is on AI and neural network performance. But there’s none the less going to be some very happy data scientists out there, especially among academics.
| AMD Server Accelerator Specification Comparison | ||||
| Radeon VII | Radeon Instinct MI50 |
Radeon Instinct MI25 |
FirePro S9170 | |
| Stream Processors | 3840 (60 CUs) |
3840 (60 CUs) |
4096 (64 CUs) |
2816 (44 CUs) |
| ROPs | 64 | 64 | 64 | 64 |
| Base Clock | 1450MHz | 1450MHz | 1400MHz | - |
| Boost Clock | 1750MHz | 1746MHz | 1500MHz | 930MHz |
| Memory Clock | 2.0Gbps HBM2 | 2.0Gbps HBM2 | 1.89Gbps HBM2 | 5Gbps GDDR5 |
| Memory Bus Width | 4096-bit | 4096-bit | 2048-bit | 512-bit |
| Half Precision | 27.6 TFLOPS | 26.8 TFLOPS | 24.6 TFLOPS | 5.2 TFLOPS |
| Single Precision | 13.8 TFLOPS | 13.4 TFLOPS | 12.3 TFLOPS | 5.2 TFLOPS |
| Double Precision | 3.5 TFLOPS (1/4 rate) |
6.7 TFLOPS (1/2 rate) |
768 GFLOPS (1/16 rate) |
2.6 TFLOPS (1/2 rate) |
| DL Performance | ? | 53.6 TFLOPS | 12.3 TFLOPS | 5.2 TFLOPS |
| VRAM | 16GB | 16GB | 16GB | 32GB |
| ECC | No | Yes (full-chip) | Yes (DRAM) | Yes (DRAM) |
| Bus Interface | PCIe Gen 3 | PCIe Gen 4 | PCIe Gen 3 | PCIe Gen 3 |
| TDP | 300W | 300W | 300W | 275W |
| GPU | Vega 20 | Vega 20 | Vega 10 | Hawaii |
| Architecture | Vega (GCN 5) |
Vega (GCN 5) |
Vega (GCN 5) |
GCN 2 |
| Manufacturing Process | TSMC 7nm | TSMC 7nm | GloFo 14nm | TSMC 28nm |
| Launch Date | 02/07/2019 | 09/2018 | 06/2017 | 07/2015 |
| Launch Price (MSRP) | $699 | - | - | $3999 |
Speaking of AI, it should be noted that machine learning performance is another area where AMD is throttling the card. Unfortunately, more details aren’t available at this time. But given the unique needs of the ML market, I wouldn’t be surprised to find that INT8/INT4 performance is held back a bit on the Radeon VII. Or for that matter certain FP16 dot products.
Also on the chopping block is full-chip ECC support. Thanks to the innate functionality of HBM2, all Vega cards already have free ECC for their DRAM. However Vega 20 takes this one step further with ECC protection for its internal caches, and this is something that the Radeon VII doesn’t get access to.
Finally, Radeon VII also cuts back a bit on Vega 20’s off-chip I/O features. Though AMD hasn’t made a big deal of it up to now, Vega 20 is actually their first PCI-Express 4.0-capable GPU, and this functionality is enabled on the Radeon Instinct cards. However for Radeon VII, this isn’t being enabled, and the card is being limited to PCIe 3.0 speeds (so future Zen 2 buyers won’t quite have a PCIe 4.0 card to pair with their new CPU). Similarly, the external Infinity Fabric links for multi-GPU support have been disabled, so the Radeon VII will only be a solo act.
On the whole, there’s nothing very surprising about AMD’s choices here, especially given Radeon VII’s target market and target price. But these are notable exclusions that are going to matter to certain users. And if not to drive those users towards a Radeon Instinct, then they’re sure to drive those users towards the inevitable Vega 20-powered Radeon Pro.
Facebook
Twitter
RSS
289 Comments
View All Comments
PeachNCream - Thursday, February 7, 2019 - link
Sorry about that. The Radeon VII is very much out of the range of prices I'm willing to pay for any single component or even an whole system for that matter. I was zinging about the GPU being called high-end (which it rightfully is) because in another recent article, a $750 monitor was referred to as midrange. See:https://www.anandtech.com/show/13926/lg-launches-3...
It was more to make a point about the inconsistency with which AT classifies products than an actual reflection of my own buying habits.
As for my primary laptop, my daily driver is a Bay Trail HP Stream 11 running Linux so yeah, it's packing 2GB of RAM and 32GB of eMMC. I have a couple other laptops around which I use significantly less often that are older, but arguably more powerful. The Stream is just a lot easier to take from place to place.
Korguz - Friday, February 8, 2019 - link
it could be.. that maybe the manufacturer refers it as a mid range product ( the monitor ) in their product stack.. and AT.. just calls it that, because of that ?:-)
eva02langley - Friday, February 8, 2019 - link
I follow you on that. I bought a 1080 TI and I told myself this is the maximum I am willing to put for a GPU.I needed something for 4k and it was the only option. If Navi is 15% faster than Vega 64 for 300$, I am buying one on launch.
D. Lister - Saturday, February 9, 2019 - link
But why would you want to spend $300 for a downgrade from your 1080Ti?HollyDOL - Thursday, February 7, 2019 - link
Purely on gaming field this can't really compete with RTX 2080 (unless some big enough perf change comes with new drivers soon)... it's performing almost same, but at a little bit more power, hotter and almost 10dB louder, which is quite a lot. Given that it won't be able to offer anything more (as oposed to possible adoptions of DXR) I would expect it not trying to compete for same price level RTX 2080 does.If it can get $50-$100 lower otoh, you get what many people asked for... kind of "GTX 2080" ... classic performance without ray tracing and DLSS extensions.
With current price though It only makes sense if they are betting they can get enough compute buyers.
Oxford Guy - Thursday, February 7, 2019 - link
Yeah, because losing your hearing to tinnitus is definitely worth that $50-100.HollyDOL - Friday, February 8, 2019 - link
Well, it's "lab conditions", it can always get dampened with good chasis or chasis position to reasonable levels and hopefully noone should be playing with head stuck inside the chasis... For me subjectively it would be too loud, but I wanted to give the card advantage of doubt, non-reference designs should hopefully get to lower levels.Oxford Guy - Friday, February 8, 2019 - link
1) The Nvidia card will be quieter in a chassis. So, that excuse fails.2) I am not seeing significant room for doubt. Fury X was a quiet product (except at idle which some complained about, and in terms of, at least in some cases, coil whine). AMD has chosen to move backward, severely, in the noise department with this product.
This card has a fancy copper vapor chamber with flattened heatpipes and three fans. It also runs hot. So, how is it, at all, rational to expect 3rd-party cards to fix the noise problem? 3rd-party makers typically use 3 slot designs to increase clocks and they typically cost even more.
HollyDOL - Friday, February 8, 2019 - link
Well, not really. If the quieter chassis cuts of enough dB to get it out of disturbing level it will be enough. Also depends on environment... If you play in loud environment (day, loud speakers) the noise won't be percieved as bad as if you play it during night with quiter speakers. Ie. what can be sufferable during day can turn in complete hell during night.That being said I am by any means not advocating +10dB, because it is a lot, but in the end it doesn't have to present so terrible obstacle.
It is very early, there can always be a bug in drivers or bios causing this temp/noise issue or it can be a design problem that cannot be circumvented. But that will be seen only after some time. I remember bug in ForceWare causing my old GTX580 not dropping to 2D frequencies once it kicked in 3D (or was it on 8800GT, I don't really remember)... You had to restart the machine. Such things simply can happen, which doesn't make them any better ofc.
Oxford Guy - Friday, February 8, 2019 - link
"If the quieter chassis cuts of enough dB to get it out of disturbing level it will be enough."Nope. I've owned the Antec P180. I have extensively modified cases and worked hard with placement to reduce noise.
Your argument that the noise can simply be eliminated by putting it into a case is completely bogus. In fact, Silent PC Review showed that more airflow, from less restriction (i.e. a less closed-in case design) can substantially reduce GPU noise — the opposite of the P180 philosophy that Silent PC Review once advocated (and helped to design).
The other problem for your argument is that it is 100% logically true that there is zero reason to purchase an inferior product. Since this GPU is not faster than a 2080 and costs the same there is zero reason to buy a louder GPU, since, in actuality, noise doesn't just get absorbed and disappear when you put it into a case. In fact, this site wrote a review of a Seasonic PSU that could be heard "from rooms away" and I can hear noisy GPUs through walls, too.
"It is very early, there can always be a bug in drivers or bios causing this temp/noise issue"
Then it shouldn't be on the market and shouldn't have been sampled. Alpha quality designs shouldn't be review subjects, particularly when they're being passed off as the full product.