NVIDIA GeForce GTX 670 vs NVIDIA RTX A2000 Mobile Comparison

NVIDIA
GEFORCE

NVIDIA GeForce GTX 670

CORE STATE GK104
VRAM 2 GB
CLOCK SPEED 980 MHz
TDP 170 W
BUS WIDTH 256 bit
ARCHITECTURE Kepler
nm
PROCESS 28 nm
LAUNCH DATE 2012
VS
NVIDIA
GEFORCE

RTX A2000 Mobile

CORE STATE GA107
VRAM 4 GB
CLOCK SPEED 1687 MHz
TDP 95 W
BUS WIDTH 128 bit
ARCHITECTURE Ampere
nm
PROCESS 8 nm
LAUNCH DATE 2021

PERFORMANCE BENCHMARKS

geekbench_metal
7,424
N/A
geekbench_opencl
15,341
56,518
geekbench_vulkan
15,553
53,146
passmark_directx_10
N/A
57
passmark_directx_11
N/A
68
passmark_directx_12
N/A
47
passmark_directx_9
N/A
115
passmark_g2d
N/A
491
passmark_g3d
N/A
9,611
passmark_gpu_compute
N/A
4,334

Analysis: NVIDIA GeForce GTX 670 vs NVIDIA RTX A2000 Mobile

NVIDIA’s RTX A2000 Mobile and GeForce GTX 670 occupy different corners of the GPU landscape, separated by nearly a decade of architecture evolution. The A2000 Mobile, an Ampere-generation part from 2021, is a mobile-class integrated GPU, while the GTX 670, a Kepler part from 2012, was a dual-slot desktop card. The recorded benchmark data shows a decisive performance gap, with the A2000 Mobile winning both available head-to-head tests by wide margins. This analysis breaks down their architectural differences, individual strengths, specification contrasts, and the exact benchmark deltas that define their relationship.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA RTX A2000 Mobile has an average benchmark score of 13821, which places it in the 55th percentile of all GPUs. The GeForce GTX 670 scores 12773 on average, putting it in the 52nd percentile.

Q: How does the RTX A2000 Mobile compare to its nearest rivals?

A: The A2000 Mobile is 0.1% ahead of the AMD Radeon 660M, 0.4% behind the AMD Radeon RX 570X, 0.6% ahead of the AMD Radeon RX 7900 XT, and 1.5% behind the NVIDIA Tesla K10.

Q: What is the performance delta in OpenCL between the two cards?

A: In the Geekbench OpenCL test, the RTX A2000 Mobile scores 56518 versus the GTX 670’s 15341, resulting in a 268.4% advantage for the A2000 Mobile.

Q: Which card has the higher memory bandwidth?

A: The two cards are nearly identical in bandwidth. The RTX A2000 Mobile delivers 192.0 GB/s over a 128-bit bus, while the GTX 670 achieves 192.3 GB/s over a 256-bit bus.

Q: What is the transistor density difference between the two chips?

A: The A2000 Mobile’s GA107 chip packs 43.5 million transistors per mm², whereas the GTX 670’s GK104 chip has a density of 12.0 million transistors per mm².

Q: Does the GTX 670 support hardware ray tracing?

A: No. The GTX 670 has no ray tracing cores or tensor cores listed in the database. The RTX A2000 Mobile includes 20 RT cores and 80 tensor cores.

Architecture Differences

The two GPUs come from entirely different design eras. The RTX A2000 Mobile uses the GA107 chip built on Samsung’s 8 nm process, with 8,700 million transistors packed into a 200 mm² die. The GeForce GTX 670 uses the GK104 chip fabricated by TSMC on a 28 nm process, containing 3,540 million transistors across a 294 mm² die. The A2000 Mobile’s transistor density of 43.5M per mm² is over 3.5 times higher than the GTX 670’s 12.0M per mm².

Architecturally, the A2000 Mobile is based on Ampere, which enables features the Kepler-based GTX 670 cannot offer. The A2000 Mobile supports DirectX 12 Ultimate (12_2), Vulkan 1.4, and includes 20 RT cores and 80 tensor cores for ray tracing and AI workloads. The GTX 670 supports DirectX 12 (11_0) and Vulkan 1.2.175, with no RT or tensor core hardware present. The A2000 Mobile also has explicit FP16 throughput rated at 1:1 with FP32, while the GTX 670 lists no FP16 capability.

The memory subsystems reflect their respective generations. The A2000 Mobile uses 4 GB of GDDR6 at 12 Gbps effective, while the GTX 670 uses 2 GB of GDDR5 at 6 Gbps. Despite the newer memory type, the A2000 Mobile’s 128-bit bus limits bandwidth to 192.0 GB/s, nearly matching the GTX 670’s 192.3 GB/s achieved via a wider 256-bit bus. The A2000 Mobile is an IGP (integrated graphics package) with no power connectors, while the GTX 670 is a dual-slot card requiring two 6-pin connectors and a 450 W suggested PSU. The A2000 Mobile draws 95 W, compared to the GTX 670’s 170 W.

Where Each One Wins

The RTX A2000 Mobile wins every recorded benchmark comparison. In the two head-to-head tests available, it leads by 268.4% in Geekbench OpenCL and 241.7% in Geekbench Vulkan. The A2000 Mobile’s benchmark suite also includes DirectX and PassMark tests where it scores 9611 in G3D, 4334 in GPU compute, and 115 in DirectX 9. These results indicate a card that dominates in raw compute throughput and modern API performance.

The GTX 670, while losing the head-to-head, still has a recorded benchmark set that shows its historical strengths. It has a Geekbench Metal score of 7424, a Geekbench OpenCL score of 15341, and a Geekbench Vulkan score of 15553. In the database’s overall percentile ranking, the GTX 670 sits at 52, only three points below the A2000 Mobile’s 55, which suggests that in legacy workloads or specific driver-optimized scenarios, the older card remains competitive relative to the broader GPU population. However, the direct comparisons show no scenario where the GTX 670 takes the lead.

For use-case analysis, the A2000 Mobile is clearly the better choice for modern workloads: ray-traced rendering, tensor-based compute, and any application leveraging DirectX 12 Ultimate features. The GTX 670’s higher texture rate (109.8 GTexel/s versus 135.0 GTexel/s for the A2000 Mobile) is the only metric where it approaches parity, but its lower pixel rate (27.44 GPixel/s versus 80.98 GPixel/s) and far lower FP32 throughput (2.634 TFLOPS versus 8.637 TFLOPS) cap its ceiling. The GTX 670’s larger physical footprint and higher power draw also make it less suitable for compact or power-constrained systems.

Specification Differences

The two GPUs differ in nearly every measurable specification. The process node moves from 28 nm (GTX 670) to 8 nm (A2000 Mobile). Transistor count rises from 3,540 million to 8,700 million, while die size shrinks from 294 mm² to 200 mm². Clock speeds: the A2000 Mobile runs at 1215 MHz base and 1687 MHz boost, versus 915 MHz base and 980 MHz boost for the GTX 670. Memory clock is 1500 MHz (12 Gbps effective) for the A2000 Mobile and 1502 MHz (6 Gbps effective) for the GTX 670.

Core counts differ substantially. The A2000 Mobile has 2560 shading units, 80 TMUs, 48 ROPs, 20 RT cores, and 80 tensor cores. The GTX 670 has 1344 shading units, 112 TMUs, and 32 ROPs, with no RT or tensor cores. The A2000 Mobile’s pixel rate is 80.98 GPixel/s versus 27.44 GPixel/s, and its texture rate is 135.0 GTexel/s versus 109.8 GTexel/s. FP32 compute is 8.637 TFLOPS for the A2000 Mobile and 2.634 TFLOPS for the GTX 670.

Other differences: TDP is 95 W (A2000 Mobile) versus 170 W (GTX 670). The A2000 Mobile uses PCIe 4.0 x16, the GTX 670 uses PCIe 3.0 x16. Display outputs are portable-device dependent for the A2000 Mobile, while the GTX 670 has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The GTX 670 has physical dimensions of 241 mm length, 111 mm height, and 38 mm width; the A2000 Mobile lists no dimensions as it is an integrated package. The GTX 670 has a launch MSRP of 399 USD; the A2000 Mobile has no launch MSRP recorded. Release dates are 2021-04-11 for the A2000 Mobile and 2012-05-09 for the GTX 670.

Head-to-Head Benchmarks

The database records two direct benchmark comparisons between these GPUs, and the RTX A2000 Mobile wins both decisively.

In Geekbench OpenCL, the A2000 Mobile scores 56518 against the GTX 670’s 15341. This is a delta of 268.4% in favor of the A2000 Mobile. The magnitude of this gap reflects the architectural leap: the A2000 Mobile has nearly 3.3 times the FP32 throughput (8.637 TFLOPS versus 2.634 TFLOPS) and over 1.9 times the shading units (2560 versus 1344). The OpenCL workload also benefits from the A2000 Mobile’s tensor cores and newer memory subsystem, despite the bandwidth being nearly identical. The GTX 670’s GDDR5 memory and Kepler compute layout cannot keep pace with the Ampere design’s parallel execution capabilities.

In Geekbench Vulkan, the A2000 Mobile scores 53146, while the GTX 670 scores 15553. The delta is 241.7% for the A2000 Mobile. Vulkan is a low-level API that exposes hardware capabilities directly, and the results show the A2000 Mobile’s modern feature set, including DirectX 12 Ultimate support and Vulkan 1.4, translates into substantially better frame throughput in compute and graphics workloads. The GTX 670’s Vulkan support (1.2.175) is functional, but the older Kepler architecture lacks the asynchronous compute and resource-management efficiencies of Ampere.

The A2000 Mobile also holds a 55th percentile ranking versus the GTX 670’s 52nd, and its average benchmark score of 13821 exceeds the GTX 670’s 12773 by roughly 8.2%. In the broader context of the database, the A2000 Mobile sits close to the AMD Radeon RX 7900 XT (0.6% ahead) and the AMD Radeon 660M (0.1% ahead), while the GTX 670 is nearly tied with the AMD FirePro W5100 (0.6% behind) and the AMD Radeon Pro 455 (0.4% behind). The performance separation between the two cards is consistent across every measured dimension, with no recorded test where the GTX 670 closes the gap.

DETAILED SPECIFICATIONS

SPECIFICATION
GTX 670
RTX A2000 Mobile
Core Specs
Shading Units
1,344
2,560 +90.5%
Shaders
1,344
2,560 +90.5%
TMUs
112
80 -28.6%
ROPs
32
48 +50.0%
SM Count
20
Clocks
Base Clock
915 MHz
1215 MHz
Boost Clock
980 MHz
1687 MHz
Memory Clock
1502 MHz 6 Gbps effective
1500 MHz 12 Gbps effective
Memory
Memory Size
2 GB
4 GB
VRAM (MB)
2,048
4,096 +100.0%
Memory Type
GDDR5
GDDR6
Memory Bus
256 bit
128 bit
Bandwidth
192.3 GB/s
192.0 GB/s
Cache
L1 Cache
16 KB (per SMX)
128 KB (per SM)
L2 Cache
512 KB
2 MB
Performance
Pixel Rate
27.44 GPixel/s
80.98 GPixel/s
Texture Rate
109.8 GTexel/s
135.0 GTexel/s
FP32 (TFLOPS)
2.634 TFLOPS
8.637 TFLOPS
FP64 (TFLOPS)
109.8 GFLOPS (1:24)
135.0 GFLOPS (1:64)
FP16 (TFLOPS)
8.637 TFLOPS (1:1)
AI/RT
RT Cores
20
Tensor Cores
80
Power
TDP
170 W
95 W
TDP (W)
170
95 -44.1%
Suggested PSU
450 W
Power Connectors
2x 6-pin
None
Architecture
Architecture
Kepler
Ampere
GPU Name
GK104
GA107
Generation
GeForce 600
Ampere-MW (Ax000)
Process Size
28 nm
8 nm
Transistors
3,540 million
8,700 million
Die Size
294 mm²
200 mm²
Foundry
TSMC
Samsung
Density
12.0M / mm²
43.5M / mm²
API Support
DirectX
12 (11_0)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.2.175
1.4
OpenCL
3.0
3.0
CUDA
3.0
8.6
Shader Model
6.5 (5.1)
6.8
Physical
Slot Width
Dual-slot
IGP
Length
241 mm 9.5 inches
Height
111 mm 4.4 inches
Outputs
2x DVI1x HDMI 1.4a1x DisplayPort 1.2
Portable Device Dependent
Bus Interface
PCIe 3.0 x16
PCIe 4.0 x16
Other
Launch Price
399 USD
Production
End-of-life
End-of-life
Predecessor
GeForce 500
Quadro Turing-M
Successor
GeForce 700
Ada-MW
View GeForce GTX 670 Details View RTX A2000 Mobile Details