Instinct MI455X vs P102-101

Primary details

GPU architecture, market segment, value for money and other general parameters compared.

Place in the ranking	not rated	not rated
Place by popularity	not in top-100	not in top-100
Architecture	Pascal (2016−2021)	CDNA 5.0 (2026)
GPU code name	GP102	MI450 256CU
Market segment	Workstation	Workstation
Release date	2018 (8 years ago)	23 July 2026 (recently)

Detailed specifications

General parameters such as number of shaders, GPU core base clock and boost clock speeds, manufacturing process, texturing and calculation speed. Real power consumption of some graphics cards can well exceed their nominal TDP, especially if overclocked.

Pipelines / CUDA cores	3200	32768
Core clock speed	1557 MHz	1000 MHz
Boost clock speed	1670 MHz	2400 MHz
Number of transistors	11,800 million	320,000 million
Manufacturing process technology	16 nm	2 nm
Power consumption (TDP)	250 Watt	2500 Watt
Texture fill rate	334.0	2,457.6
Floating-point processing power	10.69 TFLOPS	no data
ROPs	80	no data
TMUs	200	1024
L1 Cache	1.2 MB	16 MB
L2 Cache	2.5 MB	192 MB

Form factor & compatibility

Information on compatibility with other computer components. Useful when choosing a future computer configuration or upgrading an existing one. For desktop graphics cards it's interface and bus (motherboard compatibility), additional power connectors (power supply compatibility).

Interface	PCIe 3.0 x4	PCIe 6.0 x16
Length	267 mm	no data
Width	2-slot	EAM Module
Supplementary power connectors	2x 8-pin	None

VRAM capacity and type

Parameters of VRAM installed: its type, size, bus, clock and resulting bandwidth. Integrated GPUs have no dedicated video RAM and use a shared part of system RAM.

Memory type	GDDR5	HBM4
Maximum RAM amount	10 GB	432 GB
Memory bus width	320 Bit	24576 Bit
Memory clock speed	2002 MHz	1900 MHz
Memory bandwidth	320.3 GB/s	23.3 TB/s
Resizable BAR	-	+

Connectivity and outputs

This section shows the types and number of video connectors on each GPU. The data applies specifically to desktop reference models (for example, NVIDIA’s Founders Edition). OEM partners often modify both the number and types of ports. On notebook GPUs, video‐output options are determined by the laptop’s design rather than the graphics chip itself.

Display Connectors

No outputs

No outputs

API and SDK support

List of supported 3D and general-purpose computing APIs, including their specific versions.

DirectX	12 (12_1)	N/A
Shader Model	6.7	N/A
OpenGL	4.6	N/A
OpenCL	3.0	3.0
Vulkan	1.3	N/A
CUDA	6.1	-

Pros & cons summary

---

Maximum RAM amount	10 GB	432 GB
Chip lithography	16 nm	2 nm
Power consumption (TDP)	250 Watt	2500 Watt