2025-03-11

AMD Ryzen 9 9950X3D Delivers Excellent Performance For Linux Developers, Creators & Technical Computing

Ahead of tomorrow’s availability of the Ryzen 9 9900X3D and Ryzen 9 9950X3D CPUs in retail channels, today the embargo lifts on being able to deliver Ryzen 9 9950X3D reviews and performance benchmarks. Simply put, for Linux creators, developers, enthusiasts, and others running technical computing workloads and other similar tasks on their desktop, the Ryzen 9 9950X3D with its 16 cores / 32 threads and 144MB total cache makes for an excellent desktop CPU. In this review are around 400 Linux benchmarks looking at the captivating performance and competitive power efficiency of the AMD Ryzen 9 9950X3D.

Up today is the Phoronix review of the AMD Ryzen 9 9950X3D under Linux. While both the Ryzen 9 9900X3D and 9950X3D are launching tomorrow, AMD has not provided review samples to reviewers for the 12-core Ryzen 9 9900X3D processor. But I do intend to buy a Ryzen 9 9900X3D on launch day (assuming availability…) so stay tuned for AMD Ryzen 9 9900X3D Linux benchmarks/review hopefully later in the week on Phoronix. The Ryzen 9 9950X3D and 9900X3D processors follow the 8-core Ryzen 7 9800X3D that launched back in November.

The AMD Ryzen 9 9950X3D features 16 x Zen 5 cores for a total of 32 threads, features a 4.3GHz base clock, and a 5.7GHz maximum boost clock. There is a 16MB L2 cache and a total of 128MB of L3 cache thanks to AMD 3D V-Cache found on one of its two CCDs. The AMD Ryzen 9 9950X3D has a default TDP of 170 Watts. The Ryzen 9 9950X3D is unlocked for overclocking if that interests you.

This AM5 processor otherwise is quite similar to the Ryzen 9 9950X and other Ryzen 9000 series Zen 5 desktop processors but is now their top-end SKU with 3D V-Cache. The AMD Ryzen 9 9950X3D is expected to hit worldwide availability on 12 March at $699 USD compared to the Ryzen 9 9950X currently retailing for around $545 USD.

With my AMD Ryzen 9 9950X3D testing over the past few weeks it’s been working out very well under Linux. With an updated BIOS the processor works just fine in existing AM5 motherboards and with any relatively recent Linux kernel you should be in good shape. With Linux 6.13+ there is the AMD 3D V-Cache Optimizer driver available and in a separate article this week I’ll be looking at its impact for the Ryzen 9 9950X3D – in this review today all the CPUs were at their default driver settings.

The CPUs recently all (re)tested for this comparison included:

- Ryzen 7 7800X3D
- Ryzen 7 9700X
- Ryzen 7 9800X3D
- Ryzen 9 7900
- Ryzen 9 7900X
- Ryzen 9 7900X3D
- Ryzen 9 7950X
- Ryzen 9 7950X3D
- Ryzen 9 9900X
- Ryzen 9 9950X
- Ryzen 9 9950X3D
- Core Ultra 5 245K
- Core Ultra 9 285K

All the benchmarks were carried out on Ubuntu 24.10 with the Linux 6.13 kernel for a fresh upstream kernel experience.

AMD Ryzen 9 9950X3D Linux Benchmarks

From there around 400 different benchmarks were carried out for a very diverse look at the performance capabilities of the AMD Ryzen 9 9950X3D on Linux. Thanks to AMD for supplying the Ryzen 9 9950X3D review kit for today’s embargo lift to be able to provide these timely Linux performance benchmarks.

Timed Linux Kernel Compilation benchmark with settings of Build: allmodconfig. Ryzen 9 9950X3D was the fastest.

As we enjoyed with the Ryzen 7000 series, for those developers carrying out frequent code compilation or even tasking a desktop system for CI/CD build box purposes, the Ryzen 9 9950X3D allows achieving faster build times than what was otherwise achievable with the Ryzen 9 9950X. With the X3D 16-core variant, compiling the Linux kernel with all available x86_64 modules led the build time dropping from 549 to 531 seconds… Or the Intel Core Ultra 9 285K taking 577 seconds. For those carrying out many compilation jobs throughout the day or as a CI/CD box, those seconds can easily add up for big projects.

Timed Linux Kernel Compilation benchmark with settings of Build: allmodconfig. Ryzen 9 9950X3D was the fastest.

These faster Linux kernel builds were achieved with the Ryzen 9 9950X3D consuming the same amount of power as the Ryzen 9 9950X and still lower than the Intel Core Ultra 9 285K.

Timed LLVM Compilation benchmark with settings of Build System: Ninja. Ryzen 9 9950X3D was the fastest.

The Ryzen 9 9950X3D also shaved off 15 seconds of an LLVM compilation job compared to the Ryzen 9 9950X.

Timed Gem5 Compilation benchmark with settings of Time To Compile. Ryzen 9 9950X3D was the fastest.

Timed Node.js Compilation benchmark with settings of Time To Compile. Ryzen 9 9950X3D was the fastest.

Timed PHP Compilation benchmark with settings of Time To Compile. Ryzen 9 9950X3D was the fastest.

Timed Godot Game Engine Compilation benchmark with settings of Time To Compile. Ryzen 9 9950X3D was the fastest.

For anyone doing a lot of code compilation as a software developer, CI/CD testing, or even a very enthusiastic Arch Linux or Gentoo user, the Ryzen 9 9950X3D was consistently providing faster build times over the Ryzen 9 9950X. And stronger than Intel’s Core Ultra 9 285K flagship Arrow Lake desktop processor.

7-Zip compression was also faster on Linux with the Ryzen 9 9950X3D.

Stargate Digital Audio Workstation benchmark with settings of Sample Rate: 44100, Buffer Size: 512. Ryzen 9 9950X3D was the fastest.

Stargate Digital Audio Workstation benchmark with settings of Sample Rate: 44100, Buffer Size: 1024. Ryzen 9 9950X3D was the fastest.

Stargate Digital Audio Workstation benchmark with settings of Sample Rate: 96000, Buffer Size: 512. Ryzen 9 9950X3D was the fastest.

Blender benchmark with settings of Blend File: Fishy Cat, Compute: CPU-Only. Ryzen 9 9950X3D was the fastest.

For artists using the Stargate digital audio workstation software, the Ryzen 9 9950X3D was providing slightly better performance over the Ryzen 9 9950X.

Blender benchmark with settings of Blend File: Barbershop, Compute: CPU-Only. Ryzen 9 9950X3D was the fastest.

Blender benchmark with settings of Blend File: Junkshop, Compute: CPU-Only. Ryzen 9 9950X3D was the fastest.

For 3D artists, the Blender 4.3 render performance didn’t see any significant uplift with the 16-core X3D processor.

LuxCoreRender benchmark with settings of Scene: DLSC, Acceleration: CPU. Core Ultra 9 285K was the fastest.

LuxCoreRender benchmark with settings of Scene: LuxCore Benchmark, Acceleration: CPU. Ryzen 9 9950X was the fastest.

LuxCoreRender benchmark with settings of Scene: Orange Juice, Acceleration: CPU. Ryzen 9 9950X3D was the fastest.

Similarly for LuxCore you would be better off with the cheaper Ryzen 9 9950X (non-X3D) processor.

Chaos Group V-RAY benchmark with settings of Mode: CPU. Ryzen 9 9950X3D was the fastest.

IndigoBench benchmark with settings of Acceleration: CPU, Scene: Supercar. Ryzen 9 9950X3D was the fastest.

IndigoBench benchmark with settings of Acceleration: CPU, Scene: Bedroom. Ryzen 9 9950X3D was the fastest.

But for some renderers like V-RAY and IndigoBench there were benefits on Linux to using the Ryzen 9 9950X3D.

Embree benchmark with settings of Binary: Pathtracer ISPC, Model: Asian Dragon. Ryzen 9 9950X3D was the fastest.

Embree benchmark with settings of Binary: Pathtracer ISPC, Model: Crown. Ryzen 9 9950X3D was the fastest.

Intel’s Embree ray-tracing kernels performed better with the Ryzen 9 9950X3D while not consuming any extra power.

OpenFOAM benchmark with settings of Input: drivaerFastback, Small Mesh Size, Execution Time. Ryzen 9 9950X3D was the fastest.

Turning to more technical computing workloads like the OpenFOAM computational fluid dynamics software, there continues to be excellent benefits here to using AMD processors with 3D V-Cache… Very nice speed-ups over the Ryzen 9 9950X and also the prior-generation Zen 4 X3D processors.

OpenRadioss benchmark with settings of Model: Bird Strike on Windshield. Ryzen 9 9950X3D was the fastest.

OpenRadioss benchmark with settings of Model: Rubber O-Ring Seal Installation. Ryzen 9 9950X was the fastest.

OpenRadioss benchmark with settings of Model: Cell Phone Drop Test. Ryzen 9 9950X3D was the fastest.

OpenRadioss benchmark with settings of Model: Bumper Beam. Core Ultra 9 285K was the fastest.

OpenRadioss benchmark with settings of Model: INIVOL and Fluid Structure Interaction Drop Container. Ryzen 9 9950X3D was the fastest.

There were very nice gains too with Altaur’s OpenRadioss software from using the Ryzen 9 9950X3D with 3D V-Cache. In several tests the Ryzen 9 9950X3D now bumped the Core Ultra 9 285K out of the top spot.

GPAW benchmark with settings of Input: Carbon Nanotube. Ryzen 9 9950X3D was the fastest.

Similarly, with the GPAW DFT program the Ryzen 9 9950X3D now secured the top spot while consuming less power than the Core Ultra 9 285K.

Xcompact3d Incompact3d benchmark with settings of Input: input.i3d 129 Cells Per Direction. Ryzen 9 9950X3D was the fastest.

Xcompact3d Incompact3d benchmark with settings of Input: input.i3d 193 Cells Per Direction. Core Ultra 9 285K was the fastest.

Incompact3D is another workload that has benefited from AMD 3D V-Cache and it helped here too with the Ryzen 9 9950X3D.

Pennant benchmark with settings of Test: leblancbig. Ryzen 9 9950X3D was the fastest.

Pennant benchmark with settings of Test: sedovbig. Ryzen 9 9950X3D was the fastest.

NAS Parallel Benchmarks benchmark with settings of Test / Class: SP.B. Ryzen 9 9950X3D was the fastest.

GROMACS benchmark with settings of Implementation: MPI CPU, Input: water_GMX50_bare. Ryzen 9 9950X3D was the fastest.

GROMACS benchmark with settings of Input: water_GMX50_bare. Core Ultra 9 285K was the fastest.

SPECFEM3D benchmark with settings of Model: Layered Halfspace. Core Ultra 9 285K was the fastest.

SPECFEM3D benchmark with settings of Model: Water-layered Halfspace. Core Ultra 9 285K was the fastest.

Of course, it will be much more interesting for these technical computing and HPC workloads once AMD introduced any EPYC 9005 3D V-Cache processors.

miniBUDE benchmark with settings of Implementation: OpenMP, Input Deck: BM1. Ryzen 9 9950X was the fastest.

ACES DGEMM benchmark with settings of Sustained Floating-Point Rate. Ryzen 9 9950X3D was the fastest.

LAMMPS Molecular Dynamics Simulator benchmark with settings of Model: 20k Atoms. Core Ultra 9 285K was the fastest.

ASKAP benchmark with settings of Test: tConvolve MPI, Degridding. Ryzen 9 9950X3D was the fastest.

ASKAP benchmark with settings of Test: tConvolve MPI, Gridding. Ryzen 9 9950X3D was the fastest.

ASKAP benchmark with settings of Test: tConvolve MT, Degridding. Core Ultra 9 285K was the fastest.

ASKAP benchmark with settings of Test: tConvolve OpenMP, Degridding. Ryzen 9 9950X3D was the fastest.

In some workloads the Core Ultra 9 285K performed better due to simply having more physical cores when not leveraging the SMT / sibling threads in some of the HPC workloads, but overall the Ryzen 9 9950X3D was dominating across all of these technical computing workloads.

nginx benchmark with settings of Connections: 500. Core Ultra 9 285K was the fastest.

nginx benchmark with settings of Connections: 1000. Ryzen 9 9950X3D was the fastest.

With the Nginx HTTPS web server the Ryzen 9 9950X3D was able to close the gap with the Core Ultra 9 285K while having greater energy efficiency than the Arrow Lake desktop processor.

Apache Cassandra benchmark with settings of Test: Writes. Ryzen 9 9950X3D was the fastest.

PostgreSQL benchmark with settings of Scaling Factor: 100, Clients: 500, Mode: Read Only. Ryzen 9 9950X3D was the fastest.

PostgreSQL benchmark with settings of Scaling Factor: 100, Clients: 500, Mode: Read Only, Average Latency. Ryzen 9 9950X3D was the fastest.

ClickHouse benchmark with settings of 100M Rows Hits Dataset, First Run / Cold Cache. Ryzen 9 9950X3D was the fastest.

ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. Ryzen 9 9950X3D was the fastest.

For some database workloads and other SOHO server programs the Ryzen 9 9950X3D delivered additional uplift over the Ryzen 9 9950X.

SVT-AV1 benchmark with settings of Encoder Mode: Preset 13, Input: Bosphorus 4K. Ryzen 9 9950X3D was the fastest.

SVT-AV1 benchmark with settings of Encoder Mode: Preset 8, Input: Bosphorus 4K. Ryzen 9 9950X3D was the fastest.

SVT-AV1 benchmark with settings of Encoder Mode: Preset 5, Input: Bosphorus 4K. Ryzen 9 9950X3D was the fastest.

SVT-AV1 benchmark with settings of Encoder Mode: Preset 3, Input: Bosphorus 4K. Ryzen 9 9950X3D was the fastest.

Kvazaar benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Slow. Ryzen 9 9950X was the fastest.

Kvazaar benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Very Fast. Ryzen 9 9950X was the fastest.

uvg266 benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Slow. Ryzen 9 9950X was the fastest.

VVenC benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Fast. Core Ultra 9 285K was the fastest.

VVenC benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Faster. Core Ultra 9 285K was the fastest.

x265 benchmark with settings of Video Input: Bosphorus 4K. Ryzen 9 9950X3D was the fastest.

NeatBench benchmark with settings of Acceleration: CPU. Ryzen 9 9950X3D was the fastest.

The AMD Ryzen 9 9950X3D with 3D V-Cache did help out for some video encoding workloads like with the SVT-AV1 encoder but less so in some of the other open-source video encoders.

srsRAN Project benchmark with settings of Test: PUSCH Processor Benchmark, Throughput Total. Ryzen 9 9950X3D was the fastest.

srsRAN Project benchmark with settings of Test: PDSCH Processor Benchmark, Throughput Total. Ryzen 9 9950X3D was the fastest.

The AMD Ryzen 9 9950X3D provided interesting uplift for the srsRAN 5G software stack.

TensorFlow benchmark with settings of Device: CPU, Batch Size: 64, Model: ResNet-50. Ryzen 9 9950X3D was the fastest.

TensorFlow also benefited nicely from the AMD Ryzen 9 9950X3D.

OpenVINO benchmark with settings of Model: Age Gender Recognition Retail 0013 FP16-INT8, Device: CPU. Ryzen 9 9950X was the fastest.

OpenVINO benchmark with settings of Model: Person Detection FP16, Device: CPU. Ryzen 9 9950X3D was the fastest.

OpenVINO benchmark with settings of Model: Weld Porosity Detection FP16-INT8, Device: CPU. Ryzen 9 9950X was the fastest.

OpenVINO benchmark with settings of Model: Vehicle Detection FP16-INT8, Device: CPU. Ryzen 9 9950X3D was the fastest.

OpenVINO benchmark with settings of Model: Person Vehicle Bike Detection FP16, Device: CPU. Ryzen 9 9950X3D was the fastest.

OpenVINO benchmark with settings of Model: Machine Translation EN To DE FP16, Device: CPU. Ryzen 9 9950X3D was the fastest.

OpenVINO benchmark with settings of Model: Handwritten English Recognition FP16-INT8, Device: CPU. Ryzen 9 9950X3D was the fastest.

OpenVINO benchmark with settings of Model: Road Segmentation ADAS FP16-INT8, Device: CPU. Ryzen 9 9950X3D was the fastest.

OpenVINO benchmark with settings of Model: Noise Suppression Poconet-Like FP16, Device: CPU. Ryzen 9 9950X3D was the fastest.

The open-source OpenVINO AI toolkit also benefited greatly in a number of scenarios from the 3D V-Cache with the Ryzen 9 9950X3D.

Whisper.cpp benchmark with settings of Model: ggml-base.en, Input: 2016 State of the Union. Ryzen 9 9950X3D was the fastest.

Whisper.cpp benchmark with settings of Model: ggml-small.en, Input: 2016 State of the Union. Ryzen 9 9950X3D was the fastest.

Whisper.cpp was another AI workload where the Ryzen 9 9950X3D allowed for greater performance without any additional power costs.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 512. Ryzen 9 9950X3D was the fastest.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 1024. Ryzen 9 9950X3D was the fastest.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Prompt Processing 512. Ryzen 9 9950X3D was the fastest.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Mistral-7B-Instruct-v0.3-Q8_0, Test: Prompt Processing 512. Ryzen 9 9950X3D was the fastest.

The Llama.cpp AI performance was helped by the 3D V-Cache with the Ryzen 9 9950X3D processor.

ONNX Runtime benchmark with settings of Model: ResNet101_DUC_HDC-12, Device: CPU, Executor: Standard. Ryzen 9 9950X3D was the fastest.

Geometric Mean Of All Test Results benchmark with settings of Result Composite, AMD Ryzen 9 9950X3D Linux Benchmarks. Ryzen 9 9950X3D was the fastest.

For nearly 400 benchmarks here is the geometric mean across the very wide range of workloads tested. At a high level the Ryzen 9 9950X3D was slightly faster than the Ryzen 9 9950X but largely comes down to technical computing, code compilation, some creator workloads, and more where the 3D V-Cache with Zen 5 can really pay off. The Ryzen 9 9950X3D only widened the lead AMD has over the Core Ultra 9 285K Arrow Lake flagship from Intel.

Generationally the Ryzen 9 9950X3D came out 25% faster than the Ryzen 9 7950X3D overall making it still a very nice generational upgrade with being able to leverage the same AM5 motherboard and DDR5 system memory.

Here is the side-by-side of the Ryzen 9 9950X vs. 9950X3D for showing the areas where 3D V-Cache really is helpful:

AMD Ryzen 9 9950X3D Linux Benchmarks

Heavy hitting gains for the Ryzen 9 9950X3D in AI workloads like OpenVINO, TensorFlow, and Llama.cpp. HPC / technical computing workloads with ASKAP, easyWave, CloverLeaf, Incompact3D, SPECFEM3D, OpenFOAM, OpenRadioss, GROMACS, and more all benefited handsomely from 3D V-Cache. Some server workloads like the ClickHouse database and srsRAN all benefited a great deal too from the increased cache size. Beyond just gaming that is widely promoted as being a big winner with 3D V-Cache on Ryzen, there are many other workloads able to make great use of the expanded cache size on the Ryzen 9 9950X3D.

CPU Power Consumption Monitoring Overview benchmark with settings of Accumulated CPU Power Consumption Monitoring.

Making the Ryzen 9 9950X3D results even more compelling is that it had similar power consumption to the Ryzen 9 7950X and Ryzen 9 9950X3D processors. The Ryzen 9 9950X3D also had a similar CPU power consumption average to the Core Ultra 9 285K but with a much lower peak power consumption.

AMD Ryzen 9 9950X3D Linux Benchmarks

For those debating whether it’s worthwhile upgrading from a Zen 4 Ryzen 9 7950X3D to Ryzen 9 9950X3D, here is a side-by-side against those metrics…. In workloads able to leverage the full 512-bit data path of AVX-512 with Zen 5, there were some wild generational improvements from Cryptsetup to OpenVINO and more. Tons of workloads with double digit percentage improvements upgrading from the generation-old Ryzen 9 7950X3D to the new Ryzen 9 9950X3D.

The Ryzen 9 9950X3D performed very well across many different Linux workloads and showed its huge potential for creators, developers, some server workloads, and more. The Ryzen 9 9950X3D is a great desktop processor and makes me all the more excited for the next AMD server processors with 3D V-Cache. Thanks to AMD for supplying the Ryzen 9 9950X3D review sample for launch day Linux testing at Phoronix. In the coming days I hope to be able to provide Ryzen 9 9900X3D Linux benchmarks so stay tuned for those 12-core X3D numbers.