With the Radeon Pro W7700, AMD is now closing the huge gap between the previously tested Radeon Pro W7800 and W7600, thereby completing its own portfolio of RDNA3 graphics cards for productive use. Even though this chip and the corresponding design as a gaming graphics card would place it very elegantly between a Radeon RX 7600 and the RX 7700XT, the focus of today’s test is on its use in productive applications.
Unfortunately, availability in Germany is still close to zero, so of course I can’t give any street prices. But it can be assumed that even the RRP could still be just under the magic 1000 euro mark (or just over). That’s quite a statement, because you can compare the card with various counterparts from the NVIDIA cosmos. You can find out exactly where you end up in the conclusion, because you really have to differentiate fairly between areas where the Radeon Pro W7700 really performs well and those where the picture is rather mixed. Here, too, driver optimizations can certainly be made over time, but only the current status quo counts today, the rest can be checked later.
Where does AMD see this card?
The Radeon Pro W7700 is AMD’s latest addition to the Radeon PRO series of graphics cards for workstation PCs and finally fills the aforementioned large gap in its own portfolio. This card is a mid-range model of the new generation and, despite its relatively low price, promises to deliver impressive performance in the areas of AI and CAD. Equipped with the RDNA 3 architecture and 16 GB GDDR6 memory, the W7700 offers an excellent price/performance ratio, which is said to be 1.7 times better than its predecessors, according to AMD.
The Radeon Pro W7700 is specially designed for designers and creative professionals who want to improve their workflows in applications such as Solidworks, Adobe Premiere or Da Vinci Resolve. It offers up to 28 TFLOPS performance (in FP32 values) and is also well equipped for tasks involving ray tracing, with a corresponding number of ray accelerators and 48 compute units (CUs). In addition, it features AMD’s new Radiance Display Engine, which enables impressive resolutions and vivid color depths for projects on up to four displays.
As such, the Radeon Pro W7700 looks to be a strong choice for professionals looking for a powerful yet affordable graphics card for AI, CAD and other demanding applications. Its combination of advanced technology, high performance and competitive price makes it an interesting option in the world of workstation graphics cards, according to AMD. To verify this, I also tested full versions of Solidworks, Creo, Adobe Creative Cloud and other applications, but more on that in a moment.
The Radeon Pro W7700 in detail
The card in the black light metal dress with the chrome applications actually looks more massive than it really is. Ok, the blower design with the radial fan naturally helps to save weight, so that the scales ultimately only show 833 grams. For a 24.2 cm long card, that’s perfectly acceptable.
The card is a true dual-slot design without a backplate, so that the net thickness is 3.5 cm (3.8 cm with the rear screws and the clamping cross). The installation height is based on the slot panel and is a standard 10.5 cm. Power is supplied via a normal 6 2 pin connector at the end of the card, so that no cables interfere with the installation.
On the slot bracket we find the necessary air outlet openings for the DHE principle of the card as well as a total of four DisplayPorts 2.1, which is actually sufficient for all situations in the areas where the card positions itself. Of course, this also puts a little finger in NVIDIA’s wound, as the current workstation graphics cards of the Ada generation are still stuck with version 1.4 of the DisplayPort with significantly lower throughput. Although it is considered more of a mid-range card in the new Radeon PRO series, with “only” half the RAM of the W7800 and half the CUs of the W7900, the price could make it an attractive option for many professional users.
AI acceleration and AV1 codec support: The Radeon Pro W7700 offers hardware-based AI accelerators and supports the AV1 codec for efficient video encoding processes. This is especially useful for applications that benefit from AI enhancements, such as DaVinci Resolve and Adobe Premiere.
GPU Name | RDNA 3, Navi 32 (“Wheat Nas”) |
Foundry and Node | TSMC (5 nm GCD N5, 6 nm N6 MCD) |
Transistors | 28.1 billion (81.2 million per mm²) |
Die Size | 346 mm² (200 mm² GCD, 4x 36.6 mm GCD) |
Memory | 16 GB GDDR6 |
Memory bus | 256 bit (576.0 GB/s bandwidth) |
Shading units | 3072 |
TMUs/ROPs | 192 / 96 |
CUs/RT cores | 48 / 48 |
L2 / L3 cache | 2 MB / 64 MB |
Power supply | 1x 6 2 pin |
TDP | 190 Watts |
After I had tested the card in the original and “out of the box”, I of course also disassembled it thoroughly, you never know… And curiosity always wins, of course. On the next page!
17 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Veteran
Urgestein
Urgestein
Urgestein
Urgestein
Urgestein
Veteran
Urgestein
Mitglied
Urgestein
1
Urgestein
Veteran
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →