AMD Radeon 890M vs NVIDIA GRID K2 Comparison

AMD
RADEON

AMD Radeon 890M

CORE STATE Strix Point
VRAM System Shared
CLOCK SPEED 2900 MHz
TDP 15 W
BUS WIDTH System Shared
ARCHITECTURE RDNA 3.5
nm
PROCESS 4 nm
LAUNCH DATE 2024
VS
NVIDIA
GEFORCE

GRID K2

CORE STATE GK104
VRAM 4 GB
CLOCK SPEED
TDP 225 W
BUS WIDTH 256 bit
ARCHITECTURE Kepler
nm
PROCESS 28 nm
LAUNCH DATE 2013

PERFORMANCE BENCHMARKS

3dmark_3dmark_steel_nomad_dx12
590
N/A
geekbench_opencl
37,254
10,602
geekbench_vulkan
40,808
N/A
passmark_directx_10
33
N/A
passmark_directx_11
73
N/A
passmark_directx_12
37
N/A
passmark_directx_9
97
N/A
passmark_g2d
979
N/A
passmark_g3d
8,076
N/A
passmark_gpu_compute
4,157
N/A
geekbench_metal
N/A
5,557

Analysis: AMD Radeon 890M vs NVIDIA GRID K2

Head-to-Head Benchmarks

The recorded data shows a single direct head-to-head comparison between the AMD Radeon 890M and the NVIDIA GRID K2, and the result is decisively in favor of the AMD part. In the Geekbench OpenCL test, the Radeon 890M scores 37,254 points against the GRID K2's 10,602 points. That is a delta of 251.4%, meaning the AMD integrated graphics solution delivers more than three and a half times the compute throughput of NVIDIA's older accelerator. This is not a marginal victory; it is a categorical gap that defines the entire competitive landscape between these two products.

Context from the database's average benchmark scores reinforces the same conclusion. The Radeon 890M carries an average benchmark score of 9,210 across all recorded tests, placing it at the 45th percentile of all GPUs in the database. The GRID K2, by contrast, sits at 8,080 average with a 42nd percentile ranking. While the percentile gap appears modest, the OpenCL result shows how far apart the two are in raw compute workloads. The Radeon 890M's nearest rivals in the database include the AMD Radeon Vega 8 (average 9,221, a 0.1% difference) and the NVIDIA GeForce GTX 960 (average 9,273, a 0.7% difference). The GRID K2's closest competitors are the NVIDIA GeForce GTX 650 Ti Boost (average 8,067, 0.2% higher) and the GeForce 945M (average 8,099, 0.2% higher). The Radeon 890M effectively performs in the company of mid-range desktop cards from a few generations ago, while the GRID K2 lingers near entry-level mobile and older desktop parts.

The Radeon 890M also shows a broader benchmark footprint in the database. It has recorded scores across multiple test suites: 3DMark Steel Nomad DX12 at 590, Geekbench Vulkan at 40,808, Passmark G2D at 979, Passmark G3D at 8,076, and Passmark GPU Compute at 4,157. The GRID K2 has only two recorded benchmarks, Geekbench OpenCL at 10,602 and Geekbench Metal at 5,557. The absence of DirectX and Vulkan entries for the GRID K2 in the database is itself telling, as it limits the ability to compare the two across gaming-oriented workloads. Where data exists, the AMD part wins outright.

Architecture Differences

The two GPUs come from fundamentally different eras and design philosophies. The AMD Radeon 890M is built on the RDNA 3.5 architecture, using the Strix Point chip, and is classified in the Navi III IGP generation for Strix Point Mobile. It is fabricated on a 4 nm process at TSMC, with 34,000 million transistors packed into a 233 mm² die. That yields a transistor density of 145.9 million transistors per square millimeter. The NVIDIA GRID K2 uses the Kepler architecture with the GK104 chip, manufactured on a 28 nm process, also at TSMC. It contains 3,540 million transistors on a 294 mm² die, for a density of 12.0 million transistors per square millimeter. The Radeon 890M therefore packs over 12 times more transistors per square millimeter, a direct consequence of the much newer fabrication node.

Core configuration also diverges sharply. The Radeon 890M has 1,024 shading units, 64 texture mapping units, 32 raster operation units, and 16 ray tracing cores. The GRID K2 has 1,536 shading units and 128 TMUs, but shares the same 32 ROP count. The NVIDIA part has no ray tracing cores and no tensor cores, while the AMD part also lists no tensor cores but does include dedicated RT hardware. The Radeon 890M's higher clock behavior, with a base of 400 MHz and boost of 2,900 MHz, helps it overcome the GRID K2's raw unit count advantage. The GRID K2's clocks are not recorded in the database, so a direct clock comparison is not possible, but the throughput figures speak for themselves.

Memory architecture is another major divider. The Radeon 890M uses system shared memory, with the bus width, type, and capacity all listed as system dependent. Its bandwidth is similarly dependent on the host system. The GRID K2 uses 4 GB of dedicated GDDR5 memory on a 256 bit bus, delivering 160.0 GB/s of bandwidth. The GRID K2's memory clock is listed at 1250 MHz with 5 Gbps effective. For the Radeon 890M, memory performance varies with the platform, which is an important caveat for anyone evaluating the AMD part. The Radeon 890M's pixel rate is 92.80 GPixel/s and texture rate is 185.6 GTexel/s, while the GRID K2 manages 23.84 GPixel/s and 95.36 GTexel/s. The AMD part is roughly four times faster in pixel throughput and nearly twice as fast in texture throughput.

Feature support also differs by generation. The Radeon 890M supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GRID K2 is limited to DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Radeon 890M also has modern hardware ray tracing, which the GRID K2 completely lacks. The GRID K2 is a dual-slot card with 1x 6-pin and 1x 8-pin power connectors and a suggested power supply of 550 W. The Radeon 890M is an IGP with no power connectors and a 15 W TDP, compared to the GRID K2's 225 W TDP. The GRID K2 is 267 mm long (10.5 inches), while the Radeon 890M has no recorded dimensions since it is integrated into a mobile processor.

The Verdict

The data points to a clear conclusion: the AMD Radeon 890M is the superior computing part in every measured category. Its Geekbench OpenCL score is 251.4% higher than the GRID K2's, and its average benchmark score of 9,210 exceeds the GRID K2's 8,080. The Radeon 890M also holds a higher percentile ranking, 45th versus 42nd. The architectural advantages, including a 4 nm process, ray tracing cores, and modern API support, further separate the two.

The GRID K2 does retain one advantage: its dedicated 4 GB GDDR5 memory with 160.0 GB/s bandwidth. For workloads that depend on dedicated VRAM rather than system shared memory, the GRID K2 may behave differently, though the database does not contain benchmark data to confirm this. The GRID K2 also has more shading units (1,536 versus 1,024) and more TMUs (128 versus 64), but the Radeon 890M's much higher clocks and architectural efficiency overcome this deficit in the recorded results. The GRID K2 is an end-of-life product from 2013, while the Radeon 890M is an active product released in 2024. For any new deployment, the Radeon 890M is the only sensible choice based on the recorded data.

FAQ

Q: How much faster is the AMD Radeon 890M than the NVIDIA GRID K2 in Geekbench OpenCL?

A: The Radeon 890M scores 37,254 compared to the GRID K2's 10,602, which is a 251.4% difference.

Q: Which GPU has a higher average benchmark score?

A: The Radeon 890M has an average benchmark score of 9,210, while the GRID K2 has 8,080. The Radeon 890M also ranks at the 45th percentile of all GPUs, versus 42nd for the GRID K2.

Q: Does the GRID K2 have ray tracing support?

A: No. The GRID K2 has no ray tracing cores, while the Radeon 890M includes 16 ray tracing cores.

Q: What memory configurations do the two GPUs use?

A: The Radeon 890M uses system shared memory with system dependent bandwidth. The GRID K2 has 4 GB of dedicated GDDR5 memory on a 256 bit bus with 160.0 GB/s bandwidth.

Q: What are the power requirements for each?

A: The Radeon 890M has a 15 W TDP and no power connectors. The GRID K2 has a 225 W TDP, requires 1x 6-pin and 1x 8-pin power connectors, and a suggested 550 W power supply.

Q: Which GPU supports newer graphics APIs?

A: The Radeon 890M supports DirectX 12 Ultimate (12_2) and Vulkan 1.4. The GRID K2 supports DirectX 12 (11_0) and Vulkan 1.2.175.

Where Each One Wins

The AMD Radeon 890M wins in compute throughput, as demonstrated by the OpenCL benchmark. It also wins in pixel fill rate (92.80 GPixel/s versus 23.84 GPixel/s) and texture fill rate (185.6 GTexel/s versus 95.36 GTexel/s). Its FP32 performance is 5.939 TFLOPS, more than double the GRID K2's 2.289 TFLOPS. The Radeon 890M also supports modern APIs, including DirectX 12 Ultimate and Vulkan 1.4, plus hardware ray tracing. It operates at a 15 W TDP with no external power connectors, making it suitable for mobile and low-power platforms. The Radeon 890M is an active product with a 2024 release date.

The NVIDIA GRID K2 wins in memory configuration. It has 4 GB of dedicated GDDR5 memory with 160.0 GB/s bandwidth, which is not dependent on the host system. Its 1,536 shading units and 128 TMUs are numerically higher than the Radeon 890M's 1,024 shading units and 64 TMUs, though this does not translate into a benchmark win in the recorded data. The GRID K2 also has a PCIe 3.0 x16 interface, compared to the Radeon 890M's PCIe 4.0 x8. The GRID K2 has a dedicated dual-slot form factor with its own cooling solution, which may be preferable in server environments where the host processor lacks integrated graphics capabilities. Its launch MSRP was 5,199 USD.

Specification Differences

The two GPUs differ in process node, transistor count, die size, transistor density, shading units, TMUs, ROPs, pixel rate, texture rate, FP32 performance, TDP, slot width, power connectors, bus interface, API support, display outputs, production status, release date, and memory size. The Radeon 890M uses a 4 nm process with 34,000 million transistors on a 233 mm² die at 145.9 million transistors per square millimeter. The GRID K2 uses a 28 nm process with 3,540 million transistors on a 294 mm² die at 12.0 million transistors per square millimeter. The Radeon 890M has 1,024 shading units, 64 TMUs, and 32 ROPs, with a pixel rate of 92.80 GPixel/s and a texture rate of 185.6 GTexel/s. The GRID K2 has 1,536 shading units, 128 TMUs, and 32 ROPs, with a pixel rate of 23.84 GPixel/s and a texture rate of 95.36 GTexel/s. FP32 performance is 5.939 TFLOPS for the Radeon 890M and 2.289 TFLOPS for the GRID K2. The Radeon 890M has a 15 W TDP, is an IGP with no power connectors, and uses PCIe 4.0 x8. The GRID K2 has a 225 W TDP, is dual-slot with 1x 6-pin and 1x 8-pin power connectors, requires a 550 W power supply, and uses PCIe 3.0 x16. The Radeon 890M supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GRID K2 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Radeon 890M has 16 ray tracing cores; the GRID K2 has none. Memory is system shared for the Radeon 890M, while the GRID K2 has 4 GB GDDR5 on a 256 bit bus with 160.0 GB/s bandwidth. The Radeon 890M was released in 2024 and is active; the GRID K2 was released in 2013 and is end-of-life.

DETAILED SPECIFICATIONS

SPECIFICATION
890M
GRID K2
Core Specs
Shading Units
1,024
1,536 +50.0%
Shaders
1,024
1,536 +50.0%
TMUs
64
128 +100.0%
ROPs
32
32 0.0%
Compute Units
16
Clocks
Base Clock
400 MHz
Boost Clock
2900 MHz
GPU Clock
745 MHz
Memory Clock
System Shared
1250 MHz 5 Gbps effective
Memory
Memory Size
System Shared
4 GB
VRAM (MB)
4,096
Memory Type
System Shared
GDDR5
Memory Bus
System Shared
256 bit
Bandwidth
System Dependent
160.0 GB/s
Cache
L1 Cache
128 KB per Array
16 KB (per SMX)
L2 Cache
2 MB
512 KB
L0 Cache
32 KB per WGP
Performance
Pixel Rate
92.80 GPixel/s
23.84 GPixel/s
Texture Rate
185.6 GTexel/s
95.36 GTexel/s
FP32 (TFLOPS)
5.939 TFLOPS
2.289 TFLOPS
FP64 (TFLOPS)
371.2 GFLOPS (1:16)
95.36 GFLOPS (1:24)
FP16 (TFLOPS)
5.939 TFLOPS (1:1)
AI/RT
RT Cores
16
Power
TDP
15 W
225 W
TDP (W)
15
225 +1400.0%
Suggested PSU
550 W
Power Connectors
None
1x 6-pin + 1x 8-pin
Architecture
Architecture
RDNA 3.5
Kepler
GPU Name
Strix Point
GK104
Generation
Navi III IGP (Strix Point Mobile)
GRID (K2)
Process Size
4 nm
28 nm
Transistors
34,000 million
3,540 million
Die Size
233 mm²
294 mm²
Foundry
TSMC
TSMC
Density
145.9M / mm²
12.0M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 (11_0)
OpenGL
4.6
4.6
Vulkan
1.4
1.2.175
OpenCL
2.1
3.0
CUDA
3.0
Shader Model
6.8
6.5 (5.1)
Physical
Slot Width
IGP
Dual-slot
Length
267 mm 10.5 inches
Outputs
Portable Device Dependent
No outputs
Bus Interface
PCIe 4.0 x8
PCIe 3.0 x16
Other
Launch Price
5,199 USD
Production
Active
End-of-life
Predecessor
Navi II IGP
View Radeon 890M Details View GRID K2 Details