AMD Instinct MI300 Could Feature Quad MCM GPUs Based on CDNA 3 Architecture
AMD is about to launch its Instinct MI200 GPU accelerator which will be the first chip to feature an MCM graphics architecture but it looks like the next-generation Instinct MI300, featuring the CDNA 3 architecture, is going to blow it away with a quad MCM design.
AMD Instinct MI300 Rumored To Feature Four MCM GPUs Based on CDNA 3 Graphics Architecture
The AMD Instinct MI200 accelerator with its CDNA 2 architecture is expected to feature two GPU dies on the same package. These will be connected by an Infinity Fabric interconnect and will also feature a separate die that will serve as a multi-tier cache between the two GPUs. Each graphics die will be known as GCD while the cache-die will be known as MCD.
MI300 will feature 4 GCDs 🧐
— Kepler (@Kepler_L2) September 7, 2021
There will be two CDNA 2 GPUs onboard the Instinct MI200 package but the next-generation HPC accelerator is rumored to double that. According to Kepler_L2, the Instinct MI300 will feature a 4-GCD design based on the brand new CDNA 3 architecture. The upcoming Instinct MI200 was going to feature 128 compute units per die but that has changed to 110 compute units since last week's rumor. A total of 220 Compute Units would net 14,080 cores and if we take the exact number and multiply it by 4 (the number of GCDs on Instinct MI300), we end up with 440 Compute Units or an insane 28,160 cores.
A recent AMD ROCm Developer Tools update that was spotted by Komachi did confirm a maximum of 4 MCM GPUs but those are simply 'Aldebaran' SKUs. There are expected to be at least four CDNA 2 powered Instinct accelerators with their respective (unique IDs) listed below. Note that the number doesn't represent the number of dies on each device but rather the device itself:
Now that would be true if AMD makes no changes whatsoever when moving from CDNA 2 to CDNA 3 but that's not the case. CDNA 3 is expected to bring forward a revised new architecture that won't be another Vega derivative like Arcturus or Aldebaran. The GPU architecture may also use a layout that might end up looking similar to the new WGP/SE arrangement on the new RDNA 3 chips or an entirely new design tailored towards the HPC segment. But one thing is for sure, those quad-MCM GPUs definitely are something that we can't wait to see in action!
AMD Radeon Instinct Accelerators 2020
|Accelerator Name||AMD Radeon Instinct MI6||AMD Radeon Instinct MI8||AMD Radeon Instinct MI25||AMD Radeon Instinct MI50||AMD Radeon Instinct MI60||AMD Instinct MI100||AMD Instinct MI200||AMD Instinct MI300|
|GPU Architecture||Polaris 10||Fiji XT||Vega 10||Vega 20||Vega 20||Arcturus (CDNA 1)||Aldebaran (CDNA 2)||TBA (CDNA 3)|
|GPU Process Node||14nm FinFET||28nm||14nm FinFET||7nm FinFET||7nm FinFET||7nm FinFET||Advanced Process Node||Advanced Process Node|
|GPU Dies||1 (Monolithic)||1 (Monolithic)||1 (Monolithic)||1 (Monolithic)||1 (Monolithic)||1 (Monolithic)||2 (MCM)||4 (MCM)?|
|GPU Clock Speed||1237 MHz||1000 MHz||1500 MHz||1725 MHz||1800 MHz||~1500 MHz||TBA||TBA|
|FP16 Compute||5.7 TFLOPs||8.2 TFLOPs||24.6 TFLOPs||26.5 TFLOPs||29.5 TFLOPs||185 TFLOPs||TBA||TBA|
|FP32 Compute||5.7 TFLOPs||8.2 TFLOPs||12.3 TFLOPs||13.3 TFLOPs||14.7 TFLOPs||23.1 TFLOPs||TBA||TBA|
|FP64 Compute||384 GFLOPs||512 GFLOPs||768 GFLOPs||6.6 TFLOPs||7.4 TFLOPs||11.5 TFLOPs||TBA||TBA|
|VRAM||16 GB GDDR5||4 GB HBM1||16 GB HBM2||16 GB HBM2||32 GB HBM2||32 GB HBM2||64/128 GB HBM2e?||TBA|
|Memory Clock||1750 MHz||500 MHz||945 MHz||1000 MHz||1000 MHz||1200 MHz||TBA||TBA|
|Memory Bus||256-bit bus||4096-bit bus||2048-bit bus||4096-bit bus||4096-bit bus||4096-bit bus||8192-bit||TBA|
|Memory Bandwidth||224 GB/s||512 GB/s||484 GB/s||1 TB/s||1 TB/s||1.23 TB/s||~2 TB/s?||TBA|
|Form Factor||Single Slot, Full Length||Dual Slot, Half Length||Dual Slot, Full Length||Dual Slot, Full Length||Dual Slot, Full Length||Dual Slot, Full Length||Dual Slot, Full Length / OAM||TBA|
|Cooling||Passive Cooling||Passive Cooling||Passive Cooling||Passive Cooling||Passive Cooling||Passive Cooling||Passive Cooling||TBA|