COMPARE.

Intel

Intel UHD Graphics 615

GPU

Process node

14 nm^*

Reported GPU frequency

300 MHz^*

Chip/package TDP

15 W^*

Geekbench GPU (reported compilation) · version unreported · OpenCL3,237^*Platform unreported

GFXBench 5.0 · 1080p Manhattan Offscreen40.67^* fpsWindows / OpenGL / 1920 x 1080 / offscreen · archived median

KUNLUNXIN

Kunlunxin P800

Accelerators^*

Architecture

XPU-P^*

Process node

7 nm^*

Compute processor

Kunlun 3^*

Device memory type

HBM3^*

KUNLUNXIN

Kunlunxin R200

Accelerators^*

Rated memory bandwidth

512 GB/s^*

Architecture

XPU-R^*

Process node

7 nm^*

Compute processor

Kunlun II^*

KUNLUNXIN

Kunlunxin R300

Accelerators^*

Rated memory bandwidth

512 GB/s^*

Architecture

XPU-R^*

Compute processor

Kunlun II^*

Declared accelerator TDP

150 W^*

Matrox

Matrox LUMA A310FP

GPU^*

G3D Mark · version unreported · g3d_mark3,093^* pointsPlatform unreported

No Blender rendering score

Matrox

Matrox LUMA A380P

GPU^*

G3D Mark · version unreported · g3d_mark5,811^* pointsPlatform unreported

No Blender rendering score

MATROX

Matrox LUMA Pro A310 Quad

GPU^*

Video memory

4 GB^*

Rated memory bandwidth

124 GB/s^*

GPU cores

768^*

Process node

6 nm^*

No Blender rendering score

MATROX

Matrox LUMA Pro A380 Octal

GPU^*

Video memory

12 GB^*

Process node

6 nm^*

GPUs per card

2^*

Video memory type

GDDR6^*

No Blender rendering score

MATROX

Matrox LUMA Pro A380 Quad

GPU^*

Video memory

6 GB^*

Rated memory bandwidth

186 GB/s^*

GPU cores

1,024^*

Process node

6 nm^*

No Blender rendering score

METAX

MetaX C500 OAM

Accelerators^*

Rated memory bandwidth

1,800 GB/s^*

Architecture

Xiyun^*

Process node

7 nm^*

Compute processor

MXC500^*

METAX

MetaX C500 PCIe

Accelerators^*

Rated memory bandwidth

1,800 GB/s^*

Architecture

Xiyun^*

Process node

7 nm^*

Compute processor

MXC500^*

METAX

MetaX C600

Accelerators^*

Rated memory bandwidth

3,600 GB/s^*

Architecture

Xiyun^*

Process node

7 nm^*

Compute processor

MXC600^*

MOORE THREADS

Moore Threads MTT S10

GPU^*

Video memory

2 GB^*

Rated memory bandwidth

34.13 GB/s^*

GPU cores

1,024^*

Process node

12 nm^*

No Blender rendering score

MOORE THREADS

Moore Threads MTT S1000M

Accelerators^*

Rated memory bandwidth

68.26 GB/s^*

Process node

12 nm^*

Memory bus

128 bit^*

Texture mapping units

64^*

MOORE THREADS

Moore Threads MTT S2000

Accelerators^*

Rated memory bandwidth

136.50 GB/s^*

Process node

12 nm^*

Memory bus

256 bit^*

GPU L2 cache

4 MB^*

Moore Threads

Moore Threads MTT S30

GPU^*

Video memory

4 GB^*

Rated memory bandwidth

68.26 GB/s^*

GPU cores

1,024^*

Process node

12 nm^*

G3D Mark · version unreported · g3d_mark580^* pointsPlatform unreported

No Blender rendering score

MOORE THREADS

Moore Threads MTT S3000E

Accelerators^*

Rated memory bandwidth

448 GB/s^*

Process node

7 nm^*

Memory bus

256 bit^*

Transistors

22,000 million^*

MOORE THREADS

Moore Threads MTT S4000

Accelerators^*

Rated memory bandwidth

768 GB/s^*

Device memory

48 GB^*

Host interface

PCIe 5.0 x16^*

MOORE THREADS

Moore Threads MTT S50

GPU^*

Video memory

8 GB^*

Rated memory bandwidth

136.50 GB/s^*

GPU cores

2,048^*

Process node

12 nm^*

No Blender rendering score

MOORE THREADS

Moore Threads MTT S60

GPU^*

Video memory

8 GB^*

Rated memory bandwidth

136.50 GB/s^*

GPU cores

2,048^*

Process node

12 nm^*

No Blender rendering score

Moore Threads

Moore Threads MTT S70

GPU^*

Video memory

7 GB^*

Rated memory bandwidth

392 GB/s^*

GPU cores

3,584^*

Process node

7 nm^*

G3D Mark · version unreported · g3d_mark1,915^* pointsPlatform unreported

No Blender rendering score

Moore Threads

Moore Threads S3000 MTvGPU-1132

GPU^*

G3D Mark · version unreported · g3d_mark1,997^* pointsPlatform unreported

No Blender rendering score

National Semiconductor

National Semiconductor NS32081D-10

Accelerators^*

National Semiconductor

National Semiconductor NS32081D-15

Accelerators^*