AMD Radeon PRO W7800 & W7900 Workstation Efficiency Overview – Techgage

AMD’s newest technology Radeon PROs have simply landed, and are available to us within the type of the high-end Radeon PRO W7800, and even higher-end Radeon PRO W7900. This latest sequence is constructed on AMD’s RDNA3 structure, which brings quite a few efficiency enhancements and have updates.

One of many largest adjustments on the {hardware} stage to RDNA3’s high-end components is the adoption of a chiplet design, which we’ve beforehand seen from the corporate’s Ryzen and EPYC-based processors, and in addition its RDNA3-based Radeon gaming sequence.

A chiplet design makes use of smaller dies to construct the entire chip, which makes it simpler to configure merchandise for various efficiency targets, and in addition reduces the quantity of loss from defects, as greater dies improve the chance of encountering them.

Architectural enhancements from RDNA3 that translate into noticeable perks embrace elevated ray traced, AI, and video encode efficiency, AV1 encode help, increased reminiscence bandwidth, and an up to date show engine which provides DisplayPort 2.1 to help the beefiest resolutions (8K60 uncompressed; as much as 12K60 w/ DSC).

Whereas AMD’s gaming-focused Radeon lineup tends to function a wide-range of fashions, AMD has saved issues easy on the workstation aspect. The final technology RDNA2 sequence featured solely three fashions, whereas the RDNA technology had solely two. The “hold it easy” method isn’t a foul one, however with the present lowest-end RDNA3 PRO card being priced at $2,499, we hope to see the mid-range taken care of in some unspecified time in the future down the street.

The Radeon PRO W7900 is the primary 9-series mannequin since 2017’s WX 9100 that AMD has launched, because the RDNA2 technology peaked with the W6800, and RDNA1, W5700. That implies that the brand new second-from-the-top W7800 is the efficient successor to the W6800, whereas the W7900 is launched to focus on the extremely high-end.

One factor that’s attention-grabbing in regards to the W7800 is that there isn’t any comparable card to be discovered on the gaming aspect of Radeon. So far, AMD has launched the RX 7900 XTX, RX 7900 XT, and most just lately, the RX 7600. It’s not usually that we see a efficiency hole like that in a lineup, or fashions discovered on the Radeon PRO aspect that aren’t discovered on the gaming aspect.

When this Radeon PRO sequence was first introduced, AMD talked about what it appears at when it builds new workstation GPUs, with worth/perf being an enormous concern, in addition to safety and reliability continuously being improved and optimized. With these specific W7800 and W7900 Radeon PROs, AMD is catering to those that demand a excessive stage of efficiency for essentially the most demanding workloads.

It’s spectacular to take a look at the W7800, with its big 32GB framebuffer, and notice that it’s not even the top-end providing right here. With the W7900, AMD took an enormous framebuffer and made it huge, packing 50% extra reminiscence in. AMD’s Radeon Software program exhibits the W7900 as having 49,136 MB of whole out there reminiscence (enabling ECC cuts it right down to 46,064 MB).

The framebuffers on each of those new Radeon PROs are so giant, that the majority of our conventional benchmark checks received’t have the ability to take full benefit of them. For individuals who don’t need VRAM to be the limiting issue, GPUs just like the Radeon PRO W7800 and W7900 provide an enormous quantity of freedom.

General, AMD targets these Radeon PROs at any workload and state of affairs that will profit not simply from a lot of reminiscence, however merely a sooner GPU. It talks about 4K and 8K 12-bit workloads with the likes of Adobe Premiere Professional, After Results, and DaVinci Resolve; on the sport engine and rendering aspect, it talks of Autodesk’s 3ds Max and Maya, Maxon’s Cinema 4D and Redshift, Avid Media Composer, and extra.

On the subject of Redshift, we had hoped that we’d have the ability to embrace a efficiency take a look at that renderer on this article, as Maxon has simply launched the primary public beta that features Radeon help. With this launch onslaught over, we’re planning to dig in subsequent week.

AMD additionally touts the efficiency prowess of those new Radeon PROs for design and manufacturing, which may embrace instruments like 3ds Max, D5 Render, and SolidWorks Visualize. As a testomony to the truth that AMD’s made a notable influence within the workstation market, D5 Render, SolidWorks Visualize, and Maxon Redshift had been all NVIDIA-exclusive till just lately.

Different perfect meant workloads for these new PRO playing cards embrace structure and engineering, with AMD as soon as once more speaking about Lumion, an answer we hope to have the ability to take a look at out down-the-road. After all, past these workloads talked about, any that require giant framebuffers are naturally going to really feel snug with these GPUs.

Earlier than transferring onto a take a look at efficiency, we wished to attract consideration to AMD’s Radeon Software program: PRO Version, included with the driving force. This software program has seen many updates over the previous few years, and is value exploring no matter your meant workloads or workflows. Inside it, you’ll be able to simply document and stream the desktop, regulate performance-related choices, allow hotkeys and overlays, and extra.

Within the specific part seen above, there’s an choice to allow Radeon PRO Viewport Increase, which is able to enhance the efficiency of the interactive viewport in supported functions, in addition to In-Viewport Picture Increase to extend the crispness.

Whereas it’s a function which will cater to benchmarkers greater than the everyday consumer, we’ve come to understand the overlay function, permitting us to observe GPU clock speeds and temperatures whereas working benchmarks or stress checks:

Thankfully, you probably have completely no real interest in profiting from Radeon Software program, you’ll be able to set up the driving force with out it. That mentioned, it appears precious sufficient to us to maintain it put in, since you’re certain to search out one thing in there that can profit you.

With all of that lined, let’s transfer right into a take a look at efficiency. We’re going to start with a take a look at viewport efficiency in CATIA, SolidWorks, Creo, and Siemens NX on this web page, and proceed on with viewport efficiency for 3ds Max, Maya, Power, and Medical on web page two. On the third web page, we’ll discover encoding and rendering, and at last, the fourth web page will cowl rendering checks and our conclusions.

Additionally, please notice that not each GPU listed within the testbed desk under goes to be present in each single graph, largely as a result of there are too many GPU outcomes to moderately match into one. If the chart you’re is lacking RTX 3060 Ti ~ RTX 3090 or RX 6600 XT, the end result will probably be present in one of many aforementioned articles.

There’s an infinite quantity of knowledge forward. For every of the eight viewport checks we’ve run, we’re offering each the general rating generated by SPECviewperf, in addition to a breakdown of the subtests, and their respective body charges. These charts are cramped, and are supposed to be considered at their full 1080p decision. The GPU positions don’t change between these charts, so it can hopefully be straightforward sufficient to check one GPU to a different.

Lastly, it could seem odd {that a} Turing-based NVIDIA Quadro or two are included in these outcomes, however these fashions are essentially the most present that now we have. We’re together with them largely so the efficiency advantages anticipated out of NVIDIA’s workstation driver will be higher understood. The GeForce RTX 4090 is the closest mannequin hardware-wise to NVIDIA’s current-gen RTX 6000 Ada Technology, but it surely lacks the optimizations discovered within the skilled driver.

Viewport Efficiency

PRO card or not, AMD’s Radeons dominate this CATIA take a look at, taking excessive eight positions. When the last-gen Radeon PRO W6800 beats out NVIDIA’s top-end GeForce RTX 4090, which is technically superior, it actually highlights how a lot driver optimization issues.

Once we take a look at the subtest outcomes, we are able to get a good higher glimpse of how firms will prioritize one efficiency angle over one other. AMD excels on the augmented Shaded modes, leaping far forward of the competitors, whereas NVIDIA accomplishes the identical with the usual Shaded mode. Typically efficiency variations are ridiculously particular, which is why it’s a good suggestion to take a look at the subtest breakdown reasonably than rely solely on the take a look at’s general rating.

CATIA and SolidWorks may share the identical developer, however the efficiency traits of every are fairly completely different. AMD dominated the CATIA take a look at, however NVIDIA delivers a harder battle right here. With the W7900 leaping thus far forward of the 7900 XTX, this software program is an effective instance of the place the pro-level driver optimizations show helpful.

Wanting on the subtests, we are able to see that each AMD and NVIDIA ship related ranges of efficiency more often than not, primarily based on equal {hardware}. The entire subtests noticed enchancment on the W7900 vs. the gaming-focused RX 7900 XTX. When evaluating W7900 to W7800, the top-end mannequin affords notable beneficial properties in each the Shaded and RealView modes. Be aware that whereas gaming GPUs can make the most of RealView by way of the SPECviewperf viewset, an precise workstation GPU shall be required to make use of it within the native utility.

As we noticed with the CATIA take a look at, AMD dominates the highest of our Creo chart. Primarily based on the comparisons now we have right here, we are able to see that the skilled driver advantages each AMD’s and NVIDIA’s GPUs – however particularly the latter. When a two-generation-old Turing-based RTX 6000 pulls forward of the current-gen top-end RTX 4090, it proves that drivers have a lot to do with Creo efficiency on the NVIDIA aspect.

The subtests as soon as once more assist us perceive what’s occurring under-the-hood. The complete examined GeForce lineup performs near the identical across-the-board, whereas the older Quadros leap forward within the Shaded checks. As talked about, AMD’s skilled driver additionally boosts efficiency, however to a comparably modest extent. With robust Radeon efficiency right here, the last-gen and mid-range RX 6600 XT outperforms the RTX 4090 in all however the Submarine Reflections + AO take a look at.

Siemens NX requires a workstation-class GPU for the very best expertise on Intel or NVIDIA, and whereas that was the case for Radeon, it’s not. General, AMD once more performs nice right here, though we’d count on a good battle from NVIDIA with a more moderen workstation GPU, because the two-generation-old RTX 6000 sits simply behind the Radeon PRO W7800.

We haven’t talked a lot about Intel up thus far, however with a subtest graph like this, it’s arduous to disregard the elephant within the room. Really, an elephant may in all probability fill the chasm Intel’s created there. The identical outcomes present us that whereas the NVIDIA workstation driver improves efficiency throughout all the viewport modes, it’s solely the Superior Studio mode that will ship appropriate efficiency on GeForce.

As talked about earlier, the subsequent web page will check out viewport efficiency in 3ds Max, Maya, Power, and Medical. Web page three will deal with encode checks, whereas the fourth and last web page will cowl rendering and our conclusions.