PantheonSim / GPUs / NVIDIA RTX PRO 6000 Blackwell Server Edition

Simulated NVIDIA RTX PRO 6000 Blackwell Server Edition: a GPU profile for CUDA and ROCm testing

PantheonSim, the GPU simulator for CUDA and ROCm, simulates the NVIDIA RTX PRO 6000 Blackwell Server Edition (Blackwell, 12.0) as the profile nvidia/rtx-pro-6000-server. The Server Edition of the RTX PRO 6000 Blackwell, a 96 GB GB202 card. It is the part AWS sells as its g7e instances. Its profile is verified, which means it was read from a physical card.

The profile at a glance

These are the numbers a program can ask this simulated GPU, and the values nvidia-smi show for it. They come from the profile file in the simulator's repository, pinned to one commit of it.

Property	Value
Profile id	`nvidia/rtx-pro-6000-server`
Vendor and architecture	NVIDIA, Blackwell
Compute capability	`12.0`
Device memory	95.0 GiB, GDDR
Streaming multiprocessors	188
Warp size	32
Max threads per multiprocessor	1536
Shared memory per block (opt-in maximum)	99 KiB
L2 cache	128 MiB
Memory bus width	512-bit
Power limit (profile)	600 W
Maximum clocks (profile)	2430 MHz shader, 12481 MHz memory
ECC memory	yes
PCIe link reported	Gen5 x16
Tensor or matrix cores	yes
Verified (measured on a physical card)	yes

What this GPU is for

The Server Edition of the RTX PRO 6000 Blackwell, a 96 GB GB202 card. It is the part AWS sells as its g7e instances.

Blackwell is split across several compute capabilities in these profiles: 10.0 for the B200 and GB200, 10.3 for the B300, 11.0 for Jetson Thor, 12.0 for the GeForce RTX 50 series and the RTX PRO 6000 cards, and 12.1 for the GB10. Only the RTX PRO 6000 Server Edition was read from a physical card.

Among the 37 NVIDIA profiles it ranks 11th by device memory (95.0 GiB) and 2nd by SMs (188), counting the largest as first.

Is this profile verified?

Yes. This profile is marked verified, which means its values were read from a physical NVIDIA RTX PRO 6000 Blackwell Server Edition by the simulator's characterization tools, not copied from a data sheet. Verified describes the profile's numbers: it says nothing about speed, because the simulator has no timing model.

Simulate this GPU

Build the simulator (see the docs) or open the playground, then:

nvcc -cudart shared -arch=compute_120 my_app.cu -o my_app
vgpu run --gpu nvidia/rtx-pro-6000-server ./my_app
vgpu shell -y --gpu nvidia/rtx-pro-6000-server --count 4
vgpu info --gpu nvidia/rtx-pro-6000-server

This profile is compute capability 12.0. A kernel built for compute_75 runs on it and on every other NVIDIA profile; a kernel built for a newer architecture is refused by a profile that is older than it. Blackwell targets need CUDA 12.8 or newer. Link the CUDA runtime shared (-cudart shared) so the simulator can stand in for it.

To use it every time, put gpu: nvidia/rtx-pro-6000-server in a pantheonsim.yaml settings file, or in a GitHub Actions job give the PantheonSim action gpu: nvidia/rtx-pro-6000-server (see CUDA and HIP tests in GitHub Actions). The simulator checks behaviour and has no timing model, so a run on a physical GPU before a release still counts. The CUDA guide explains the rest.

Where the numbers come from

Each profile file begins with a note on its sources. This is the note for NVIDIA RTX PRO 6000 Blackwell Server Edition, as written in the file:

Source note for nvidia/rtx-pro-6000-server

verified: true, read from an AWS g7e.2xlarge on 2026-10-09 (driver 595.91.07, CUDA 13.2) with nvidia/tools/characterize.cu: the whole characterization, and the simulator reproduces its 512 conformance values (ptx_semantics, control_flow). Corrected by it: the boost clock (2430 MHz, not the Workstation Edition's 2617), the memory clock (12481), the persisting L2 (83886080, five sixths of the L2, not three quarters), the card's CUDA memory total (101975851008 bytes, 97252 MiB; nvidia-smi's total is 97887 MiB). Public sources for the rest are listed below. Where each value comes from: - compute capability 12.0: NVIDIA's CUDA GPUs table (developer.nvidia.com/cuda-gpus); the same GB202 die as nvidia/rtx-pro-6000. - SMs 188: NVIDIA RTX Blackwell GPU Architecture whitepaper (v1.1), Table 1 (RTX PRO 6000 Blackwell Workstation Edition: 188 SMs, 24064 CUDA Cores); NVIDIA's Server Edition page gives the same 24,064 cores. This is the part AWS sells as the g7e instances ("RTX PRO Server 6000"). - memory 96 GB GDDR7 with ECC on a 512-bit interface, 1597 GB/s, up to 600 W (configurable), PCIe Gen 5: NVIDIA's RTX PRO 6000 Blackwell Server Edition page. NVIDIA's "GB" for graphics memory is binary (see rtx2080ti.yaml), so vram_bytes is 98304 MiB; a real card's nvidia-smi total is a little smaller, so this is an upper bound. The Server Edition's bandwidth (1597 GB/s) is below the Workstation Edition's 1792 GB/s: its memory runs slower. - SM clock: NVIDIA's Server Edition page gives none; the Workstation Edition's 2617 MHz (whitepaper) stands in until measured. - PCI device 0x2BB5, "NVIDIA RTX PRO 6000 Blackwell Server Edition", from the name table in NVIDIA's open GPU kernel modules (g_nv_name_released.h, main at ca2d03e, 2026-10-07). - Per-SM limits are compute capability 12.0's, the same as nvidia/rtx-pro-6000 and nvidia/rtx5090.

The GPU index lists every profile, and this one is measured on a physical card.

Other Blackwell profiles: