NVIDIA GeForce RTX 4090 vs NVIDIA RTX A6000 Comparison

NVIDIA
GEFORCE

NVIDIA GeForce RTX 4090

CORE STATE AD102
VRAM 24 GB
CLOCK SPEED 2520 MHz
TDP 450 W
BUS WIDTH 384 bit
ARCHITECTURE Ada Lovelace
nm
PROCESS 5 nm
LAUNCH DATE 2022
VS
NVIDIA
GEFORCE

RTX A6000

CORE STATE GA102
VRAM 48 GB
CLOCK SPEED 1800 MHz
TDP 300 W
BUS WIDTH 384 bit
ARCHITECTURE Ampere
nm
PROCESS 8 nm
LAUNCH DATE 2020

PERFORMANCE BENCHMARKS

3dmark_3dmark_steel_nomad_dx12
9,223
N/A
geekbench_opencl
255,416
193,937
geekbench_vulkan
271,631
164,462
passmark_directx_10
224
155
passmark_directx_11
326
191
passmark_directx_12
150
87
passmark_directx_9
397
245
passmark_g2d
1,299
913
passmark_g3d
38,194
22,577
passmark_gpu_compute
26,613
14,110

Analysis: NVIDIA GeForce RTX 4090 vs NVIDIA RTX A6000

The Verdict

The recorded data draws a clear line between these two NVIDIA offerings. The GeForce RTX 4090 wins all nine head-to-head benchmark comparisons, with deltas ranging from 31.7% to 88.6%. Its average benchmark score of 60,347 places it at the 88th percentile of all GPUs, while the RTX A6000 sits at 44,075 and the 84th percentile. The RTX 4090 is the decisive performance leader in every measured category, from compute workloads to legacy DirectX tests.

However, the RTX A6000 is not without purpose. Its 48 GB memory capacity is double that of the RTX 4090, and its workstation-oriented design with a dual-slot footprint and 8-pin EPS power connector targets a different class of system. The data suggests two distinct buyers: those who need maximum frame rates and raw throughput should choose the RTX 4090, while those whose work demands the larger memory pool and professional form factor should consider the A6000, accepting its lower performance scores. The RTX 4090 carries a launch MSRP of 1,599 USD, while the RTX A6000 launched at 4,649 USD.

Architecture Differences

The two cards represent different generations of NVIDIA GPU design. The GeForce RTX 4090 uses the AD102 chip built on the Ada Lovelace architecture, fabricated by TSMC on a 5 nm process. The RTX A6000 uses the GA102 chip with the older Ampere architecture, produced by Samsung on an 8 nm node. This process gap is visible in the transistor data: the RTX 4090 packs 76,300 million transistors into a 609 mm² die, yielding a density of 125.3 million transistors per square millimeter. The A6000 contains 28,300 million transistors on a slightly larger 628 mm² die, for a density of 45.1 million per square millimeter. The newer process allows the RTX 4090 to nearly triple the transistor count on a comparable die area.

Compute resources follow the same pattern. The RTX 4090 has 16,384 shading units, 512 texture mapping units, 176 render output units, 128 ray tracing cores, and 512 tensor cores. The A6000 has 10,752 shading units, 336 TMUs, 112 ROPs, 84 RT cores, and 336 tensor cores. Clock speeds also favor the newer card: the RTX 4090 runs at a 2235 MHz base and 2520 MHz boost, while the A6000 runs at 1410 MHz base and 1800 MHz boost. These architectural differences compound into large performance gaps in the benchmark data.

Memory architecture differs in capacity and type, though both use a 384 bit bus. The RTX 4090 pairs 24 GB of GDDR6X with a memory clock of 1313 MHz and 21 Gbps effective speed, achieving 1.01 TB/s bandwidth. The A6000 offers 48 GB of GDDR6 at 2000 MHz and 16 Gbps effective, for 768.0 GB/s bandwidth. The A6000 trades bandwidth and memory type for double the capacity.

Head-to-Head Benchmarks

The RTX 4090 dominates every recorded comparison. The largest margin appears in Passmark GPU Compute, where the RTX 4090 scores 26,613 against the A6000's 14,110, a lead of 88.6%. This gap reflects the raw compute throughput difference: the RTX 4090 is rated at 82.58 TFLOPS for both FP32 and FP16, while the A6000 is rated at 38.71 TFLOPS for both. The RTX 4090 also wins Passmark DirectX 12 by 72.4% (150 versus 87) and Passmark DirectX 11 by 70.7% (326 versus 191). Passmark G3D shows a 69.2% advantage (38,194 versus 22,577).

The smallest recorded gap is in Geekbench OpenCL, where the RTX 4090 scores 255,416 against 193,937, a 31.7% lead. Geekbench Vulkan shows a larger 65.2% margin (271,631 versus 164,462). Legacy API tests follow the trend: Passmark DirectX 9 favors the RTX 4090 by 62% (397 versus 245), Passmark DirectX 10 by 44.5% (224 versus 155), and Passmark G2D by 42.3% (1,299 versus 913). Notably, the A6000 does not win a single head-to-head test, and it lacks a recorded 3DMark Steel Nomad DX12 score that the RTX 4090 has at 9,223.

FAQ

Q: Which card has higher raw compute performance?

A: The RTX 4090 is rated at 82.58 TFLOPS for both FP32 and FP16, more than double the A6000's 38.71 TFLOPS. This shows up in the 88.6% Passmark GPU Compute lead.

Q: Does the RTX A6000 have any advantage in memory?

A: Yes, it offers 48 GB of GDDR6, twice the RTX 4090's 24 GB. However, the RTX 4090 achieves higher bandwidth at 1.01 TB/s versus 768.0 GB/s, because of its faster GDDR6X memory at 21 Gbps effective.

Q: How do their average benchmark scores compare?

A: The RTX 4090 averages 60,347 across all recorded benchmarks, placing it at the 88th percentile of all GPUs. The A6000 averages 44,075, placing it at the 84th percentile. The nearest rivals to the RTX 4090 include the Intel Arc Pro A60 at 60,326 (0% delta) and AMD Radeon Pro W6600M at 61,896 (-2.5% delta). The A6000's nearest rival is the NVIDIA GeForce RTX 4090 Mobile at 43,667 (0.9% delta).

Q: What are the physical and power differences?

A: The RTX 4090 is a triple-slot card measuring 304 mm long, 137 mm tall, and 61 mm wide, with a 450 W TDP, one 16-pin connector, and a suggested 850 W PSU. The A6000 is dual-slot, 267 mm long and 112 mm tall, with a 300 W TDP, an 8-pin EPS connector, and a suggested 700 W PSU.

Q: Which card supports more display outputs?

A: The RTX 4090 provides 1x HDMI 2.1 and 3x DisplayPort 1.4a. The A6000 provides 4x DisplayPort 1.4a and no HDMI output.

Q: Are both cards still in production?

A: No, both are listed as end-of-life. The RTX 4090 was released on 2022-09-19, and the A6000 on 2020-10-04. The RTX 4090's predecessor is the GeForce 30 series and its successor is the GeForce 50 series. The A6000's predecessor is Quadro Turing and its successor is Workstation Ada.

Where Each One Wins

The RTX 4090 wins every benchmark category in the database. Its strongest relative performance is in compute-heavy workloads: the 88.6% Passmark GPU Compute lead and the 65.2% Geekbench Vulkan gap indicate that applications leveraging general-purpose GPU compute will see the largest benefit. DirectX workloads also favor it heavily, with 72.4% and 70.7% leads in DirectX 12 and 11 respectively. The 42.3% G2D lead shows even 2D operations run faster, though the absolute scores (1,299 versus 913) are modest.

The A6000's case rests on capacity and form factor, not speed. Its 48 GB memory pool is the standout feature, double the RTX 4090's 24 GB. For workloads that exceed 24 GB of working set, the A6000 is the only one of the two that can proceed at all, regardless of the RTX 4090's speed advantage. Its dual-slot profile and 300 W TDP also make it easier to integrate into dense workstation chassis, and the 8-pin EPS connector aligns with server power delivery standards. The A6000's 4x DisplayPort 1.4a outputs suit multi-monitor professional setups, whereas the RTX 4090 splits its outputs between HDMI and DisplayPort.

Users should weigh the performance data against the memory requirement. If the workload fits within 24 GB, the RTX 4090 is faster in every measured test, often by large margins. If the workload needs more than 24 GB, the A6000 is the only viable option between these two, and its lower scores become secondary to its memory capacity.

Specification Differences

The two cards differ in nearly every hardware category. The RTX 4090 uses the AD102 chip on TSMC 5 nm, while the A6000 uses GA102 on Samsung 8 nm. Transistor counts are 76,300 million versus 28,300 million, and die sizes are 609 mm² versus 628 mm². Transistor densities are 125.3M per mm² versus 45.1M per mm².

Clock speeds favor the RTX 4090: base 2235 MHz versus 1410 MHz, boost 2520 MHz versus 1800 MHz. Memory differs in size (24 GB versus 48 GB), type (GDDR6X versus GDDR6), and effective speed (21 Gbps versus 16 Gbps). Bandwidth is 1.01 TB/s versus 768.0 GB/s.

Compute resources are higher on the RTX 4090: 16,384 shading units versus 10,752, 512 TMUs versus 336, 176 ROPs versus 112, 128 RT cores versus 84, and 512 tensor cores versus 336. Pixel rate is 443.5 GPixel/s versus 201.6 GPixel/s, and texture rate is 1,290.2 GTexel/s versus 604.8 GTexel/s. FP32 and FP16 are both 82.58 TFLOPS versus 38.71 TFLOPS.

Power and physical design differ: 450 W versus 300 W TDP, triple-slot versus dual-slot, 1x 16-pin versus 8-pin EPS, and suggested PSU of 850 W versus 700 W. Dimensions are 304 mm by 137 mm by 61 mm versus 267 mm by 112 mm with no recorded width. Display outputs are 1x HDMI 2.1 plus 3x DisplayPort 1.4a versus 4x DisplayPort 1.4a.

Both cards share several specifications: PCIe 4.0 x16 interface, DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Release dates differ by about two years: 2022-09-19 for the RTX 4090 and 2020-10-04 for the A6000. The RTX 4090's launch MSRP was 1,599 USD; the A6000's was 4,649 USD. Both are end-of-life, with the RTX 4090 succeeding the GeForce 30 series and preceding the GeForce 50 series, while the A6000 succeeds Quadro Turing and precedes Workstation Ada.

DETAILED SPECIFICATIONS

SPECIFICATION
RTX 4090
RTX A6000
Core Specs
Shading Units
16,384
10,752 -34.4%
Shaders
16,384
10,752 -34.4%
TMUs
512
336 -34.4%
ROPs
176
112 -36.4%
SM Count
128
84 -34.4%
Clocks
Base Clock
2235 MHz
1410 MHz
Boost Clock
2520 MHz
1800 MHz
Memory Clock
1313 MHz 21 Gbps effective
2000 MHz 16 Gbps effective
Memory
Memory Size
24 GB
48 GB
VRAM (MB)
24,576
49,152 +100.0%
Memory Type
GDDR6X
GDDR6
Memory Bus
384 bit
384 bit
Bandwidth
1.01 TB/s
768.0 GB/s
Cache
L1 Cache
128 KB (per SM)
128 KB (per SM)
L2 Cache
72 MB
6 MB
Performance
Pixel Rate
443.5 GPixel/s
201.6 GPixel/s
Texture Rate
1,290.2 GTexel/s
604.8 GTexel/s
FP32 (TFLOPS)
82.58 TFLOPS
38.71 TFLOPS
FP64 (TFLOPS)
1,290.2 GFLOPS (1:64)
604.8 GFLOPS (1:64)
FP16 (TFLOPS)
82.58 TFLOPS (1:1)
38.71 TFLOPS (1:1)
AI/RT
RT Cores
128
84 -34.4%
Tensor Cores
512
336 -34.4%
Power
TDP
450 W
300 W
TDP (W)
450
300 -33.3%
Suggested PSU
850 W
700 W
Power Connectors
1x 16-pin
8-pin EPS
Architecture
Architecture
Ada Lovelace
Ampere
GPU Name
AD102
GA102
Generation
GeForce 40
Workstation Ampere (Ax000)
Process Size
5 nm
8 nm
Transistors
76,300 million
28,300 million
Die Size
609 mm²
628 mm²
Foundry
TSMC
Samsung
Density
125.3M / mm²
45.1M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
8.9
8.6
Shader Model
6.8
6.8
Physical
Slot Width
Triple-slot
Dual-slot
Length
304 mm 12 inches
267 mm 10.5 inches
Height
137 mm 5.4 inches
112 mm 4.4 inches
Outputs
1x HDMI 2.13x DisplayPort 1.4a
4x DisplayPort 1.4a
Bus Interface
PCIe 4.0 x16
PCIe 4.0 x16
Other
Launch Price
1,599 USD
4,649 USD
Production
End-of-life
End-of-life
Predecessor
GeForce 30
Quadro Turing
Successor
GeForce 50
Workstation Ada
View GeForce RTX 4090 Details View RTX A6000 Details