Last update: June 19, 2026 — synced with the official NVIDIA® CUDA™ GPUs list.
Understanding CUDA™ Compute Capability
The Compute Capability identifies the architecture and feature set of an NVIDIA® GPU. It is written as a major.minor version (e.g. 8.9) and describes what the GPU can do — supported instructions, data types, and hardware features — not how fast it is. Each new GPU generation introduces a new compute capability.
When compiling CUDA™ code, you need to tell the compiler which compute capability to target. You may have encountered this error:
nvcc fatal : Unsupported gpu architecture 'compute_XX'
This usually means one of two things:
- You specified a compute capability that does not match your GPU, or
- Your installed CUDA™ Toolkit is too old to support your GPU. Support for each compute capability is added in a specific toolkit version (for example,
compute_120requires CUDA™ 12.8 or newer). With a brand-new GPU, the fix is often to upgrade the CUDA™ Toolkit, not to change the flag.
The Solution
Add the appropriate flag to your nvcc compiler command:
-gencode arch=compute_XX,code=[sm_XX,compute_XX]
Replace XX with your GPU’s compute capability value (without the decimal point — e.g. 8.9 becomes 89).
Virtual vs. Real Architectures
The -gencode flag references two kinds of architecture:
compute_XXis a virtual architecture: it produces PTX, an intermediate assembly that the driver can just-in-time (JIT) compile at runtime.sm_XXis a real architecture: it produces SASS, the actual binary code that runs on a GPU with that exact compute capability.
So arch=compute_89,code=[sm_89,compute_89] embeds both the ready-to-run SASS for compute capability 8.9 and the PTX. Keeping the PTX (code=compute_89) lets the binary JIT-compile and run on future GPUs that did not exist when you built it — at the cost of a small startup delay the first time.
Simpler Shortcuts
For common cases you do not need the full -gencode syntax:
-arch=sm_89 # target a single GPU (compute capability 8.9)
-arch=native # auto-detect the GPU of the machine you are building on
-arch=all-major # build for every major architecture your toolkit supports
-arch=native is especially handy for local builds, while -arch=all-major is a good default for portable binaries.
Finding Your GPU’s Compute Capability
Using CUDA™ Device Query
If CUDA™ is already installed on your system, use the built-in device query tool:
/usr/local/cuda/bin/__nvcc_device_query
The output is the compute capability encoded as major × 10 + minor — a two- or three-digit number depending on the GPU. For example, with a GeForce RTX 4070:
89
This corresponds to compute capability 8.9 (as shown in the architecture table below). Newer GPUs print three digits, e.g. 110 for a Jetson™ AGX Thor (11.0) or 120 for a GeForce RTX 5090 (12.0).
Using nvidia-smi
On recent drivers you can query the compute capability directly, with no compilation required:
nvidia-smi --query-gpu=compute_cap --format=csv
You can also build and run the deviceQuery program from the CUDA™ samples, which prints the compute capability together with other device details.
Online Reference
NVIDIA® maintains an official list of all GPU compute capabilities on their CUDA™ GPUs webpage.
Example: Desktop GPU
If you have an NVIDIA® GeForce RTX 4090:
- Locate it in the “Workstation/Consumer” column of the table
- Find its compute capability: 8.9
- Use the following flag:
-gencode arch=compute_89,code=[sm_89,compute_89]
Example: Embedded GPU
If you’re working with an NVIDIA® Jetson™ AGX Thor:
- Look in the “Jetson” column of the table and search for “Jetson™ T5000”
- The Jetson™ T5000 and T4000 modules have compute capability: 11.0
- Use this configuration:
-gencode arch=compute_110,code=[sm_110,compute_110]
Multi-GPU Support
For applications that need to run on various systems with different GPUs, you can compile for multiple compute capabilities:
ARCH="-gencode arch=compute_60,code=sm_60 \
-gencode arch=compute_70,code=sm_70 \
-gencode arch=compute_75,code=sm_75 \
-gencode arch=compute_80,code=sm_80 \
-gencode arch=compute_89,code=sm_89" \
nvcc ${ARCH} [other nvcc options] <input_file> -o <output_file>
This approach ensures your application is compatible across different GPU architectures, though it will increase compilation time and binary size.
Compute Capability Overview
Here’s a summary of NVIDIA® GPU architectures and their compute capabilities
Recent Architectures
| Compute Capability | Data Center | Workstation/Consumer | Jetson™ |
|---|---|---|---|
| 12.1 | NVIDIA® GB10 (DGX Spark) | ||
| 12.0 | NVIDIA® RTX PRO 6000 Blackwell Server Edition NVIDIA® RTX PRO 4500 Blackwell Server Edition |
NVIDIA® RTX PRO 6000 Blackwell Workstation Edition NVIDIA® RTX PRO 6000 Blackwell Max-Q Workstation Edition NVIDIA® RTX PRO 5000 Blackwell NVIDIA® RTX PRO 4500 Blackwell NVIDIA® RTX PRO 4000 Blackwell NVIDIA® RTX PRO 4000 Blackwell SFF Edition NVIDIA® RTX PRO 2000 Blackwell GeForce RTX 5090 GeForce RTX 5080 GeForce RTX 5070 Ti GeForce RTX 5070 GeForce RTX 5060 Ti GeForce RTX 5060 GeForce RTX 5050 |
|
| 11.0 | Jetson™ T5000 (AGX Thor) Jetson™ T4000 (Thor) |
||
| 10.3 | NVIDIA® GB300 NVIDIA® B300 |
||
| 10.0 | NVIDIA® GB200 NVIDIA® B200 |
||
| 9.0 | NVIDIA® GH200 NVIDIA® H200 NVIDIA® H100 |
||
| 8.9 | NVIDIA® L4 NVIDIA® L40 NVIDIA® L40S |
NVIDIA® RTX 6000 Ada NVIDIA® RTX 5000 Ada NVIDIA® RTX 4500 Ada NVIDIA® RTX 4000 Ada NVIDIA® RTX 4000 SFF Ada NVIDIA® RTX 2000 Ada GeForce RTX 4090 GeForce RTX 4080 GeForce RTX 4070 Ti GeForce RTX 4070 GeForce RTX 4060 Ti GeForce RTX 4060 GeForce RTX 4050 |
|
| 8.7 | Jetson™ AGX Orin Jetson™ Orin NX Jetson™ Orin Nano |
||
| 8.6 | NVIDIA® A40 NVIDIA® A10 NVIDIA® A16 NVIDIA® A2 |
NVIDIA® RTX A6000 NVIDIA® RTX A5000 NVIDIA® RTX A4000 NVIDIA® RTX A3000 NVIDIA® RTX A2000 GeForce RTX 3090 Ti GeForce RTX 3090 GeForce RTX 3080 Ti GeForce RTX 3080 GeForce RTX 3070 Ti GeForce RTX 3070 GeForce RTX 3060 Ti GeForce RTX 3060 GeForce RTX 3050 Ti GeForce RTX 3050 |
|
| 8.0 | NVIDIA® A100 NVIDIA® A30 |
||
| 7.5 | NVIDIA® T4 | QUADRO RTX 8000 QUADRO RTX 6000 QUADRO RTX 5000 QUADRO RTX 4000 QUADRO RTX 3000 QUADRO T2000 NVIDIA® T1200 NVIDIA® T1000 NVIDIA® T600 NVIDIA® T500 NVIDIA® T400 GeForce GTX 1650 Ti NVIDIA® TITAN RTX GeForce RTX 2080 Ti GeForce RTX 2080 GeForce RTX 2070 GeForce RTX 2060 |
Older Architectures
| Compute Capability | Data Center | Workstation/Consumer | Jetson™ |
|---|---|---|---|
| 7.2 | Jetson™ AGX Xavier Jetson™ Xavier NX |
||
| 7.0 | NVIDIA® V100 | Quadro GV100 NVIDIA® TITAN V |
|
| 6.2 | Jetson™ TX2 | ||
| 6.1 | Tesla P40 Tesla P4 |
Quadro P6000 Quadro P5200 Quadro P5000 Quadro P4200 Quadro P4000 Quadro P3200 Quadro P3000 Quadro P2200 Quadro P2000 Quadro P1000 Quadro P620 Quadro P600 Quadro P500 Quadro P400 P620 P520 NVIDIA® TITAN Xp NVIDIA® TITAN X GeForce GTX 1080 Ti GeForce GTX 1080 GeForce GTX 1070 Ti GeForce GTX 1070 GeForce GTX 1060 GeForce GTX 1050 |
|
| 6.0 | Tesla P100 | Quadro GP100 | |
| 5.3 | Jetson™ Nano | ||
| 5.2 | Tesla M60 Tesla M40 |
Quadro M6000 24GB Quadro M6000 Quadro M5000 Quadro M4000 Quadro M2000 Quadro M5500M Quadro M2200 Quadro M620 GeForce GTX TITAN X GeForce GTX 980 Ti GeForce GTX 980 GeForce GTX 970 GeForce GTX 960 GeForce GTX 950 GeForce GTX 980M GeForce GTX 970M GeForce GTX 965M GeForce 910M |
|
| 5.0 | Quadro K2200 Quadro K1200 Quadro K620 Quadro M1200 Quadro M520 Quadro M5000M Quadro M4000M Quadro M3000M Quadro M2000M Quadro M1000M Quadro K620M Quadro M600M Quadro M500M NVIDIA® NVS 810 GeForce GTX 750 Ti GeForce GTX 750 GeForce GTX 960M GeForce GTX 950M GeForce 940M GeForce 930M GeForce GTX 850M GeForce 840M GeForce 830M |
||
| 3.7 | Tesla K80 | ||
| 3.5 | Tesla K40 Tesla K20 |
Quadro K6000 Quadro K5200 Quadro K610M Quadro K510M GeForce GTX TITAN Z GeForce GTX TITAN Black GeForce GTX TITAN GeForce GTX 780 Ti GeForce GTX 780 GeForce GT 730 GeForce GT 720 GeForce GT 705* GeForce GT 640 (GDDR5) GeForce 920M |
|
| 3.2 | Tegra K1 Jetson™ TK1 |
||
| 3.0 | Tesla K10 | Quadro K5000 Quadro K4200 Quadro K4000 Quadro K2000 Quadro K2000D Quadro K600 Quadro K420 Quadro 410 Quadro K6000M Quadro K5200M Quadro K5100M Quadro K500M Quadro K4200M Quadro K4100M Quadro K3100M Quadro K2200M Quadro K2100M Quadro K1100M NVIDIA® NVS 510 GeForce GTX 770 GeForce GTX 760 GeForce GTX 690 GeForce GTX 680 GeForce GTX 670 GeForce GTX 660 Ti GeForce GTX 660 GeForce GTX 650 Ti BOOST GeForce GTX 650 Ti GeForce GTX 650 GeForce GT 740 GeForce GTX 880M GeForce GTX 870M GeForce GTX 780M GeForce GTX 770M GeForce GTX 765M GeForce GTX 760M GeForce GTX 680MX GeForce GTX 680M GeForce GTX 675MX GeForce GTX 670MX GeForce GTX 660M GeForce GT 755M GeForce GT 750M GeForce GT 650M GeForce GT 745M GeForce GT 645M GeForce GT 740M GeForce GT 730M GeForce GT 640M GeForce GT 640M LE GeForce GT 735M GeForce GT 730M |
|
| 2.1 | NVIDIA® NVS 315 NVIDIA® NVS 310 NVS 5400M NVS 5200M NVS 4200M Quadro 2000 Quadro 2000D Quadro 600 Quadro 4000M Quadro 3000M Quadro 2000M Quadro 1000M GeForce GTX 560 Ti GeForce GTX 550 Ti GeForce GTX 460 GeForce GTS 450 GeForce GTS 450 GeForce GT 730 DDR3,128bit GeForce GT 640 (GDDR3) GeForce GT 630 GeForce GT 620 GeForce GT 610 GeForce GT 520 GeForce GT 440 GeForce GT 440 GeForce GT 430 GeForce GT 430* GeForce 820M GeForce 800M GeForce GTX 675M GeForce GTX 670M GeForce GT 635M GeForce GT 630M GeForce GT 625M GeForce GT 720M GeForce GT 620M GeForce 710M GeForce 705M GeForce 610M GeForce GTX 580M GeForce GTX 570M GeForce GTX 560M GeForce GT 555M GeForce GT 550M GeForce GT 540M GeForce GT 525M GeForce GT 520MX GeForce GT 520M GeForce GTX 485M GeForce GTX 470M GeForce GTX 460M GeForce GT 445M GeForce GT 435M GeForce GT 420M GeForce GT 415M GeForce 710M GeForce 410M |
||
| 2.0 | Tesla C2075 Tesla C2050 Tesla C2070 Tesla M2050 Tesla M2070 Tesla M2075 Tesla M2090 |
Quadro Plex 7000 Quadro 6000 Quadro 5000 Quadro 4000 Quadro 4000 for Mac Quadro 5010M Quadro 5000M GeForce GTX 590 GeForce GTX 580 GeForce GTX 570 GeForce GTX 480 GeForce GTX 470 GeForce GTX 465 GeForce GTX 480M |
|
| 1.3 | Tesla C1060 Tesla S1070 Tesla M1060 |
Quadro FX 5800 Quadro FX 4800 Quadro FX 4800 for Mac Quadro FX 3800 Quadro CX Quadro Plex 2200 D2 GeForce GTX 295 GeForce GTX 285 GeForce GTX 285 for Mac GeForce GTX 280 GeForce GTX 275 GeForce GTX 260 |
|
| 1.2 | Quadro 400 Quadro FX 380 Low Profile NVIDIA® NVS 300 Quadro FX 1800M Quadro FX 880M Quadro FX 380M NVIDIA® NVS 300 NVS 5100M NVS 3100M NVS 2100M GeForce GT 240 GeForce GT 220 GeForce 210 GeForce GTS 360M GeForce GTS 350M GeForce GT 335M GeForce GT 330M GeForce GT 325M GeForce GT 240M GeForce G210M GeForce 310M GeForce 305M |
||
| 1.1 | Quadro FX 4700 X2 Quadro FX 3700 Quadro FX 1800 Quadro FX 1700 Quadro FX 580 Quadro FX 570 Quadro FX 470 Quadro FX 380 Quadro FX 370 Quadro FX 370 Low Profile Quadro NVS 450 Quadro NVS 420 Quadro NVS 295 Quadro Plex 2100 D4 Quadro FX 3800M Quadro FX 3700M Quadro FX 3600M Quadro FX 2800M Quadro FX 2700M Quadro FX 1700M Quadro FX 1600M Quadro FX 770M Quadro FX 570M Quadro FX 370M Quadro FX 360M Quadro NVS 320M Quadro NVS 160M Quadro NVS 150M Quadro NVS 140M Quadro NVS 135M Quadro NVS 130M Quadro NVS 450 Quadro NVS 420 Quadro NVS 295 GeForce GTS 250 GeForce GTS 150 GeForce GT 130 GeForce GT 120 GeForce G100* GeForce 9800 GX2 GeForce 9800 GTX+ GeForce 9800 GTX GeForce 9600 GSO GeForce 9500 GT GeForce 8800 GTS GeForce 8800 GT GeForce 8800 GS GeForce 8600 GTS GeForce 8600 GT GeForce 8500 GT GeForce 8400 GS GeForce 9400 mGPU GeForce 9300 mGPU GeForce 8300 mGPU GeForce 8200 mGPU GeForce 8100 mGPU GeForce GTX 285M GeForce GTX 280M GeForce GTX 260M GeForce 9800M GTX GeForce 8800M GTX GeForce GTS 260M GeForce GTS 250M GeForce 9800M GT GeForce 9600M GT GeForce 8800M GTS GeForce 9800M GTS GeForce GT 230M GeForce 9700M GT GeForce 9650M GS GeForce 9700M GT GeForce 9650M GS GeForce 9600M GT GeForce 9600M GS GeForce 9500M GS GeForce 8700M GT GeForce 8600M GT GeForce 8600M GS GeForce 9500M G GeForce 9300M G GeForce 8400M GS GeForce G210M GeForce G110M GeForce 9300M GS GeForce 9200M GS GeForce 9100M G GeForce 8400M GT GeForce G105M |
||
| 1.0 | Tesla C870 Tesla D870 Tesla S870 |
Quadro FX 5600 Quadro FX 4600 Quadro Plex 2100 S4 GeForce GT 420 GeForce 8800 Ultra GeForce 8800 GTX GeForce GT 340 GeForce GT 330 GeForce GT 320 GeForce 315 GeForce 310 GeForce 9800 GT GeForce 9600 GT GeForce 9400GT |
Happy CUDA Programming!