AMD Instinct MI300X Surpasses NVIDIA RTX 4090 as Fastest GPU in Geekbench OpenCL Benchmark

More From Author

See more articles

How Much is Microsoft Net Worth in 2025?

Microsoft Net Worth in 2025 - All You Need to Know Microsoft, a stalwart in the technology industry,...

Know Everything About Mario Bros Net Worth in April...

Mario and Luigi are two brothers who became well-known after making their first appearance in the Donkey...

LATEST: Mukesh Ambani Net Worth as of February 2025

Indian business tycoon Mukesh Ambani is the managing director, chairman, and major shareholder of the conglomerate Reliance Industries,...

The first benchmarks of AMD’s Instinct MI300X 192 GB GPU have been published on the Geekbench OpenCL benchmark, highlighting the performance capabilities of AMD’s next-generation accelerator. As such, the focus of the AMD Instinct MI300X will be squarely on Data Centers for solving the sort of AI workloads that are in demand today, putting it head-to-head with NVIDIA’s Hopper AI GPU family. Best of all, the MI300X ranks better in the GPU world than almost any of our processors in the Geekbench OpenCL benchmark-led by a significant margin.

AMD Instinct MI300X

More About AMD Instinct MI300X Benchmark

Some quick specs here with the AMD Instinct MI300X featuring the widespanning CDNA 3 arch with its 153 billion transistors in a die-chaplet packaging design. The chip includes a total of 28 dies comprising 8 HBM3 packages that allow for up to 192 GB of memory, the currently largest memory capacity of all Data Center GPUs.

The Instinct MI300X comes with 304 compute units and 19,456 cores on the GPU side. This gives it an enormous 5.3 TB/s of VRAM bandwidth and 896 GB/s of Infinity Fabric interconnect speeds. so with those two combined you’re looking at a gaming performance beast. In terms of power, the MI300X has an rated 750W TDP which is almost double of NVIDIA’s RTX 4090 with 450W rated TDP.

The AMD Instinct MI300X scored 379,660 points in the Geekbench 6.3.0 OpenCL test. These benchmarks were done with an AMD EPYC setup of two 9754 processors for a total of 128 cores / 256 threads and 3TB of system memory. Meanwhile, NVIDIA‘s flagship RTX 4090 scored 319,697 points and AMD’s top gaming GPU, the 7900 XTX scored 207,354 points.

These benchmarks highlight the remarkable speed of modern Data Center and AI GPUs. However, their high power consumption, cooling requirements, and steep prices—such as the MI300X priced around $15,000 USD—make them unsuitable for standard PC platforms.

Additionally, their specialized hardware designs and lack of optimized drivers for gaming and traditional PC applications further limit their viability outside of Data Centers, AI, HPC, and Cloud servers. AMD’s roadmap includes future releases like the MI350 and MI400 series, continuing to advance the capabilities of their Instinct GPUs in specialized computing environments.

FAQs

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.

━ Related News

Featured

━ Latest News

Featured