AMD released its CDNA-based Instinct MI100 accelerator nearly two years ago now, back in November 2020, and while the third-gen CDNA architecture is nearly here, the first-gen MI100 accelerator has hit eBay... in the nude.
The AMD Instinct MI100 accelerator is based on the CDNA architecture and made on TSMC's 7nm process node, with its GPU codenamed "Arcturus" packing 7680 FP32 cores and 32GB of HBM2 memory. The card has a 300W TDP, chewing down its power from dual 8-pin power connectors -- but not regular PCIe power connectors, but rather the EPS ones that go into your motherboard.
The card was posted on eBay in non-functional order, with a price of 500 EUR (about $500 USD or so) which is a rather huge price drop from its MSRP at the time of $6400. The card doesn't work, just to be sure you know that, so this would be a collector's item. But man, in the nude the AMD Instinct MI100 accelerator is a beautiful card under that cooler.
There are a few more pictures of the AMD Instinct MI100 accelerator in the nude thanks to the eBay listing, where you can see the massive "Victor 2033" which is a DC/DC converter, and part of the VRM assembly by the looks of things (according to "ShiinaYuriChan" on Twitter).
No gaming graphics cards on either side of the fence -- AMD, or NVIDIA -- have this, but the Instinct MI100 does. We don't normally see nudes of the PCBs of these cards, but now that we do I want to see more. I'd love to see a full teardown of the card, piece by piece, covered in great detail. I'd nerd out with that, for sure.
Here's what makes the AMD Instinct MI100 accelerator tick:
- All-New AMD CDNA Architecture - Engineered to power AMD GPUs for the exascale era and at the heart of the MI100 accelerator, the AMD CDNA architecture offers exceptional performance and power efficiency
- Leading FP64 and FP32 Performance for HPC Workloads - Delivers industry-leading 11.5 TFLOPS peak FP64 performance and 23.1 TFLOPS peak FP32 performance, enabling scientists and researchers across the globe to accelerate discoveries in industries including life sciences, energy, finance, academics, government, defense, and more.
- All-New Matrix Core Technology for HPC and AI - Supercharged performance for a full range of single and mixed-precision matrix operations, such as FP32, FP16, bFloat16, Int8, and Int4, engineered to boost the convergence of HPC and AI.
- 2nd Gen AMD Infinity Fabric Technology - Instinct MI100 provides ~2x the peer-to-peer (P2P) peak I/O bandwidth over PCIe 4.0 with up to 340 GB/s of aggregate bandwidth per card with three AMD Infinity Fabric Links. In a server, MI100 GPUs can be configured with up to two fully-connected quad GPU hives, each providing up to 552 GB/s of P2P I/O bandwidth for fast data sharing.
- Ultra-Fast HBM2 Memory - Features 32GB High-bandwidth HBM2 memory at a clock rate of 1.2 GHz and delivers an ultra-high 1.23 TB/s of memory bandwidth to support large data sets and help eliminate bottlenecks in moving data in and out of memory.
- Support for Industry's Latest PCIe Gen 4.0 - Designed with the latest PCIe Gen 4.0 technology support providing up to 64GB/s peak theoretical transport data bandwidth from CPU to GPU.