GPUs

NVIDIA GeForce RTX 5090 vs Apple 60-Core M2 Ultra Full Specs

87,040 Shaders
2.41GHz
7,680 Shaders
1.4GHz
128GB GDDR77.17TB/s
Shared Memory819.2GB/sLPDDR5
··
419 TFLOPS
··
21.47 TFLOPS
Form Factor
PCIe Card
Form Factor
iGPU
TDP
2300W
TDP
Shared
Power Connectors
4x 16-Pin 12VHPWR
Power Connectors
-

Clock Speed
···
Clock Speed
··
Peak OPS
13.41 PFLOPSFP4 Tensor Sparse
Peak OPS
21.47 TFLOPSFP16
Tensor FP4
6.7 PFLOPS
FP4 Tensor Sparse
13.41 PFLOPS
-
Tensor FP8-16
3.35 PFLOPS
FP8-16 Tensor Sparse
6.7 PFLOPS
Tensor FP8-32
1.68 PFLOPS
FP8-32 Tensor Sparse
3.35 PFLOPS
-
Tensor FP16-16
1.68 PFLOPS
FP16-16 Tensor Sparse
3.35 PFLOPS
Tensor FP16-32
838 TFLOPS
FP16-32 Tensor Sparse
1.68 PFLOPS
-
BF16
419 TFLOPS
Tensor BF16
838 TFLOPS
BF16 Tensor Sparse
1.68 PFLOPS
-
Tensor TF32
419 TFLOPS
Tensor TF32
-
FP32
419 TFLOPS
FP32
21.47 TFLOPS
FP64
6.55 TFLOPS
-
Tensor INT8
3.35 POPS
-
Ray
1.27 POPS
Ray
-
Pixel Rate
423.6 GPixel/s
Pixel Rate
335.5 GPixel/s
Texture Rate
1.64 TTexel/s
Texture Rate
671 GTexel/s

Shaders
87,040 Shaders
Shaders
7,680 Shaders
TMUs
2,720 TMUs
TMUs
480 TMUs
ROPs
704 ROPs
ROPs
240 ROPs
Tensor Cores
2,720 T-Cores
Tensor Cores
-
RT Cores
680 RT-Cores
RT Cores
-
SMs
680 SMs
EUs
960 EUs

Base Clock
2.01GHz
Base Clock
-
Boost Clock
2.41GHz
Boost Clock
1.4GHz

L2 Cache
96MB shared
L2 Cache
16MB shared

128GB GDDR7
Shared MemoryLPDDR5
Memory Bus
512-bit
Memory Bus
1024-bit
Memory Speed
28GT/s
Memory Speed
6.4GT/s
Memory Bandwidth
7.17TB/s
Memory Bandwidth
819.2GB/s
ECC
No
ECC
No

TDP
2300W
TDP
Shared
Max Temp
90°C Max
Max Temp
-

Multi-Monitor
16
Multi-Monitor
8
HDCP
HDCP 2.3
HDCP
-

12x DisplayPort 2.1
4x HDMI 2.1
-

Encoder Model
3x NVENC 9
Encoder Model
4x Apple Media Engine 2
Codec
AVC (H.264)
HEVC (H.265)
AV1
-
Codec
AVC (H.264)
HEVC (H.265)
-
ProRes

Decoder Model
2x NVDEC 6
Decoder Model
2x Apple Media Engine 2
Codec
MPEG-1
MPEG-2
MPEG-4
VC-1
VP8
VP9
AVC (H.264)
HEVC (H.265)
AV1
-
Codec
-
-
-
-
-
-
AVC (H.264)
HEVC (H.265)
-
ProRes

Form Factor
PCIe Card
Form Factor
iGPU
PCIe
8.4-Slots
PCIe
-
Height
137 mm (5.39")
Width
304 mm (11.97")
Depth
42 mm (1.65")
-
Cooling
Open-Air
2x Fans
Cooling
Passive
-
Power Connectors
4x 16-Pin 12VHPWR
Power Connectors
-

Manufacturer
Manufacturer
Chip Designer
Chip Designer
Architecture
Architecture
Family
Family
Branding
GeForce RTX 2022 branding
Branding
M2 Ultra branding
Codename
NV192
Codename
G14D
Chip Variant
GB202-250-A1
Chip Variant
-
Market Segment
Desktop
Market Segment
Workstation
Release Date
Jan 6, 2025
Release Date
Jun 5, 2023

Foundry
TSMC
Foundry
TSMC
Fabrication Node
4NP
Fabrication Node
N5P
Die Size
762mm²
Die Size
971mm²
Transistor Count
92.2 Billion
Transistor Count
134 Billion
Transistor Density
121.1 MTr/mm²
Transistor Density
138 MTr/mm²