Cache-Rich: AMD Ryzen 7 5800X3D Workstation Performance Review

[PAGETITLE=AMD Ryzen 7 5800X3D Workstation Performance Review][Top Ad] Ahead of AMD's Zen launch in early 2017, enthusiasts were clamoring for exciting progress in the desktop CPU market. While we saw GPUs gain tremendously from one generation to the next, every CPU release seemed truly modest in comparison, with 5~10% per-core performance boosts, and most features catering more to the server market than the desktop one. With that original Zen launch, AMD gave people a reason to pay close attention to desktop processors again, and while it took a couple of years to deliver a proper strike back, Intel's definitely a force to be reckoned with nowadays, with its 12th-gen Core series delivering extremely competitive performance. AMD and Intel have clearly been enjoying the challenges of figuring out where to take our CPUs next. Intel's Alder Lake architecture introduced efficient cores to sit alongside performance ones, and ultimately, it's a design that's paid-off well for the company. On the AMD side, with its Zen 4 architecture still a little ways off, the company couldn't bear to wait longer before unveiling its latest technology: 3D V-Cache. Put simply, 3D V-Cache is vertically-stacked cache that allows AMD to improve densities without expanding the CPU's resulting die size. Stacking used in chip design isn't a new concept, but AMD claims to deliver the best take on it yet, with many more interconnects sitting between each stacked chip, ensuring ultra-high bandwidth and ultra-low latency. When we first learned of 3D V-Cache, it wasn't immediately clear just how quickly it'd find itself in a desktop processor. The feature first debuted in the server-bound EPYC "Milan-X" series, and in appropriate workloads, the speed boosts can be truly notable. What about on desktop? Anticipating gains will prove to be a bit trickier. https://www.youtube.com/watch?v=dSCpVhKvmCY For reasons that are obvious to us now, AMD's primary marketing for the Ryzen 7 5800X is gaming, as the additional cache delivers improvements across many games and game engines. While most of the obvious benefit will be seen at lower resolutions, even higher resolutions can show a bit of a performance bump. Because our primary focus at Techgage is on workstation performance, we felt compelled to begin our 5800X3D testing there. Unfortunately, as the performance results ahead will attest, we probably should have started with a gaming or Linux angle, because you need specific creator workloads to tap into the additional cache - something future software updates will hopefully make more common. Nonetheless, we can follow-up with additional testing later. For now, let's review AMD's current-gen Ryzen lineup, as it stands after the introduction of the 5800X3D: AMD's Ryzen & Ryzen Threadripper Lineup Cores Clock (Turbo) L3 Cache Memory IGP TDP Price Ryzen Threadripper 3990X 64 (128T) 2.9 GHz (4.3) 288MB Quad No 280W $3990 3970X 32 (64T) 3.7 GHz (4.5) 144MB Quad No 280W $1999 3960X 24 (48T) 3.8 GHz (4.5) 140MB Quad No 280W $1399 Ryzen 9 R9 5950X 16 (32T) 3.4 GHz (4.9) 64MB Dual No 105W $799 R9 5900X 12 (24T) 3.7 GHz (4.8) 64MB Dual No 105W $549 Ryzen 7 R7 5800X3D 8 (16T) 3.4 GHz (4.5) 96MB Dual No 105W $449 R7 5800X 8 (16T) 3.8 GHz (4.7) 32MB Dual No 105W $449 R7 5700X 8 (16T) 3.4 GHz (4.6) 32MB Dual No 65W $299 Ryzen 5 R5 5600X 6 (12T) 3.7 GHz (4.6) 32MB Dual No 65W $299 R5 5600 6 (12T) 3.5 GHz (4.4) 32MB Dual No 65W $199 R5 5500 6 (12T) 3.6 GHz (4.2) 16MB Dual No 65W $159 R5 4500 6 (12T) 3.6 GHz (4.1) 11MB Dual No 65W $129 Ryzen 3 R3 4100 4 (8T) 3.8 GHz (4.0) 6MB Dual No 65W $99 R3 3300X 4 (8T) 3.8 GHz (4.3) 18MB Dual No 65W $120 R3 3100 4 (8T) 3.6 GHz (3.9) 18MB Dual No 65W $99 Ryzen w/ Radeon Vega Graphics R7 5700G 8 (16T) 3.8 GHz (4.6) 16MB Dual Yes 65W $359 R5 5600G 6 (12T) 3.9 GHz (4.4) 16MB Dual Yes 65W $259 R5 4600G 6 (12T) 3.7 GHz (4.2) 11MB Dual Yes 65W $154 Note that the Threadripper chips listed here are Zen 2, not the Zen 3-based Threadripper PRO, simply because it's a series that's more aligned with the other desktop parts. Threadripper PRO, as its name suggests, breaks out of the enthusiast market, and sits on top of the Ryzen PRO series, targeted at business. We have no idea whether AMD will even continue to update regular Threadripper, but it's a series that hasn't seen an update since early 2020. AMD's current-gen 3D V-Cache technology triples the amount of available L3 cache. This represents a bump from 32MB to 96MB with the Ryzen 7 5800X3D, and a staggering jump from 256MB to 768MB on the server-bound Milan-X EPYCs. As lightly touched on above, tripling the amount of cache isn't going to automatically improve workloads across-the-board. The reason AMD touts this CPU as the world's best gaming chip highlights that. If you do use workloads that can take advantage of a lot of cache, the performance uplift will be notable. On the creator side, you will have to try hard to find benefit, as our results in this article will also help highlight. This is why we're really eager to get gaming and Linux testing done soon. This article is going to be tackling creator workloads like image manipulation, video encoding, and rendering, and will also tackle detailed synthetics with the help of SiSoftware's Sandra. On the following page, you can get up to speed with test system configurations and our methodologies; if you don't care about that, simply head on over to page 3. [PAGEBREAK][PAGETITLE=Test Methodology & Systems][Top Ad] Benchmarking a CPU may sound like a simple enough task, but in order to deliver accurate, repeatable results, strict guidelines need to be adhered to. This makes for rigorous, time-consuming testing, but we feel that the effort is worth it. This page exists so that we can be open about how we test, and give those who care about testing procedures an opportunity to review our methodology before flaming us in the comments. Here, you can see a breakdown of all of our test machines, specifics about the tests themselves, and other general information that might be useful. For testing AMD's Ryzen AM4-based CPUs, we're using ASRock's X570 TAICHI motherboard. For Intel's 12th-gen Core processors, we're using ASUS' Z690-PLUS WiFi D4. Both motherboards support DDR4, but Intel's 12th-gen can support DDR5. We're waiting on some new components to arrive (namely a better cooler and PSU), and will be upgrading that machine soon. The same DDR4 64GB kit of Corsair DOMINATOR Platinum memory is used on each of the test platforms. After resetting each EFI to its default values, we enable the primary XMP option, and leave it at that. In the past, we obsessed over whether we should disable things like ASUS' MultiCore Enhancement, but figuring out what was "best" was so headache-inducing (everyone has a different opinion), that we're sticking to default settings here, aside from the enabling of XMP. To ensure that conditions are identical from run to run, and to improve overall reliability, we use scripts to benchmark everything in an automated manner. Most tests are run three times over, while consistent tests that vary their results <1% from run-to-run are run twice. More sensitive tests are run five times. Here's the full breakdown of the test platforms: Techgage's CPU Testing Platforms. AMD AM4 Test Platform Processors AMD Ryzen 9 5950X (3.4GHz, 16C/32T) AMD Ryzen 9 5900X (3.7GHz, 12C/24T) AMD Ryzen 7 5800X3D (3.4GHz, 8C/16T) AMD Ryzen 7 5800X (3.8GHz, 8C/16T) AMD Ryzen 5 5600X (3.7GHz, 6C/12T) Motherboard ASRock X570 TAICHI CPUs tested with BIOS P4.80 (February 16, 2022) Memory Corsair VENGEANCE (CMT64GX4M4Z3600C16) 16GB x4 Operates at DDR4-3600 16-18-18 (1.35V) Graphics NVIDIA RTX 3070 (8GB; GeForce 512.15) Storage WD Blue 3D NAND 1TB (SATA 6Gbps) Power Supply Corsair RM850X (850W) Chassis Fractal Design Define C Cooling Corsair Hydro H100i PRO RGB (240mm) Et cetera Windows 11 Pro (21H2, Build 22000.613) Intel LGA1700 Test Platform Processors Intel Core i9-12900K (3.2GHz, 16C/24T) Intel Core i5-12600K (3.7GHz, 10C/16T) Motherboard ASUS Z690-PLUS WiFi D4 CPUs tested with BIOS 1304 (March 7, 2022) Memory Corsair VENGEANCE (CMT64GX4M4Z3600C16) 16GB x4 Operates at DDR4-3600 16-18-18 (1.35V) Graphics NVIDIA RTX 3070 (8GB; GeForce 512.15) Storage WD Blue 3D NAND 1TB (SATA 6Gbps) Power Supply EVGA Bronze 600B1 (600W) Chassis Corsair Crystal X570 RGB Cooling Corsair Hydro H115i PRO RGB (280mm) Et cetera Windows 11 Pro (21H2, Build 22000.613) Testing Considerations. We use an up-to-date Windows for our testing, as well as the latest chipset driver for each respective platform. In the pursuit of accurate, repeatable benchmarks, here are some basic guidelines we follow: Disruptive services are disabled; eg: Search, Cortana, User Account Control, Defender, etc. Overlays and / or other extras are not installed with the graphics driver. Vsync is disabled at the driver level (and in any tested game or application). Default (usually Balanced) power profile is used (with screen and sleep timeouts disabled). OSes are never transplanted from one machine to another. We validate system configurations before kicking off any test run. Testing doesn't begin until the PC is idle (keeps a steady minimum wattage). All tests are repeated until there is a high degree of confidence in the results. Benchmarks of modern workloads matter, so we always try to use up-to-date software. We do not use outdated data. Performance numbers seen in this article are current. Encoding Tests. Go straight to test: Adobe Lightroom Classic Adobe Premiere Pro Agisoft Metashape Blackmagic RAW Speed Test HandBrake LameXP VEGAS Pro Rendering Tests. Go straight to test: Autodesk Arnold (Maya) Blender Chaos Corona (3ds Max) Chaos V-Ray (3ds Max) Luxion KeyShot LuxMark Maxon Cinema 4D POV-Ray Synthetic Tests. Go straight to test: SiSoftware Sandra If you think there's some information lacking on this page, or you simply want clarification on anything in particular, don't hesitate to leave a...

Cache-Rich: AMD Ryzen 7 5800X3D Workstation Performance Review

Page 987 –

Rob Williams