docs
Python SDKReference

GPU Specifications

Hardware specs for all supported GPU types -- VRAM, memory type, bandwidth, interconnect, and compute performance

Quick reference for all GPU types supported by urun.

Effective VRAM available to your model is less than total VRAM — CUDA kernels, framework overhead, and memory fragmentation typically consume 1-3 GB depending on workload.

Hardware specifications

GPUSpec stringVRAMMemory TypeMemory BandwidthInterconnect
B200 SXM"b200:N"192 GBHBM3e8.0 TB/sNVLink 5 (1.8 TB/s)
H200 SXM"h200:N"141 GBHBM3e4.8 TB/sNVLink 4 (900 GB/s)
RTX PRO 6000 Blackwell SE"rtx6000:N"96 GBGDDR71.6 TB/sPCIe 5.0 (128 GB/s)
H100 SXM"h100:N"80 GBHBM33.35 TB/sNVLink 4 (900 GB/s)
A100 SXM"a100:N"80 GBHBM2e2.0 TB/sNVLink 3 (600 GB/s)
L40S"l40s:N"48 GBGDDR6864 GB/sPCIe 4.0 (64 GB/s)
A10G"a10g:N"24 GBGDDR6600 GB/sPCIe 4.0 (64 GB/s)
L4"l4:N"24 GBGDDR6300 GB/sPCIe 4.0 (64 GB/s)

The RTX PRO 6000 Blackwell Server Edition's 96 GB of VRAM fits ~80GB-class models on a single card that would otherwise need multi-GPU sharding on 48 GB parts.

Compute performance

Dense (non-sparse) tensor-core throughput. The B200's exact dense FP16/FP8/FP4 figures depend on the final shipping clocks — confirm against NVIDIA's current Blackwell datasheet before sizing on them.

GPUFP16 TFLOPSFP8 TFLOPSFP4 TFLOPSTDP
B200 SXMsee datasheetsee datasheetsee datasheet (Blackwell-only)~1000W
RTX PRO 6000 Blackwell SEsee datasheetsee datasheetsee datasheet (Blackwell)600W
H100 SXM9891,979N/A700W
H200 SXM9891,979N/A700W
A100 SXM312N/AN/A400W
L40S362733N/A350W

H200 shares the H100's GH100 compute die, so its dense FP16/FP8 throughput matches the H100; the H200's advantage is memory (141 GB HBM3e at 4.8 TB/s vs 80 GB at 3.35 TB/s), not raw FLOPS. FP4 is a Blackwell-class (B200) capability and is not available on Hopper/Ampere/Ada parts.

Specifying a GPU

The canonical way to request hardware is the gpus="type:count" string on @app.function:

@app.function(gpus="h100:8")     # 8x H100
@app.function(gpus="l40s:1")     # 1x L40S
@app.function(gpus="rtx6000:1")  # 1x RTX PRO 6000
@app.function(gpus="a100:4")     # 4x A100
@app.function(gpus="h200")       # 1x H200 (count defaults to 1)

When you pass an integer like gpus=8 without a type, urun defaults to H100.


For guidance on choosing GPUs for your workload, see Compute.

On this page