AMD Radeon R9 M395X vs NVIDIA GeForce GTX 980 Ti Comparison

AMD
RADEON

AMD Radeon R9 M395X

CORE STATE Amethyst
VRAM 8 GB
CLOCK SPEED —
TDP 75 W
BUS WIDTH 256 bit
ARCHITECTURE GCN 3.0
nm
PROCESS 28 nm
LAUNCH DATE 2015
VS
NVIDIA
GEFORCE

GeForce GTX 980 Ti

CORE STATE GM200
VRAM 6 GB
CLOCK SPEED 1076 MHz
TDP 250 W
BUS WIDTH 384 bit
ARCHITECTURE Maxwell 2.0
nm
PROCESS 28 nm
LAUNCH DATE 2015

PERFORMANCE BENCHMARKS

geekbench_metal
33,953
19,520
geekbench_opencl
17,829
43,513
3dmark_3dmark_steel_nomad_dx12
N/A
1,321
geekbench_vulkan
N/A
47,724

Analysis: AMD Radeon R9 M395X vs NVIDIA GeForce GTX 980 Ti

The comparison between the NVIDIA GeForce GTX 980 Ti and the AMD Radeon R9 M395X is a study in contrasting design philosophies. The GTX 980 Ti is a high-power, desktop-class GPU built for maximum throughput, while the R9 M395X is a power-efficient mobile solution. Benchmark data shows a clear split: the AMD card wins decisively in Metal graphics workloads, while the NVIDIA card dominates in OpenCL compute. Their overall average scores place them in the same percentile bracket, but their performance profiles could not be more different.

The Verdict

The data indicates that neither card is a universal winner; instead, the choice hinges entirely on the workload and platform context. For users operating in a Metal-centric environment, such as a macOS ecosystem, the AMD Radeon R9 M395X is the clear choice. Its Geekbench Metal score of 33,953 is a massive 74% higher than the GTX 980 Ti's 19,520, a lead that dwarfs the differences seen in other benchmarks. This is a category-defining win for the AMD part.

Conversely, for users whose primary applications leverage OpenCL or require raw compute throughput, the NVIDIA GeForce GTX 980 Ti is the only logical pick. Its OpenCL score of 43,513 is more than double the R9 M395X's 17,829, representing a 144.1% advantage. The GTX 980 Ti also holds a lead in the overall average benchmark score, posting 28,020 versus 25,891 for the AMD card, a difference of about 8%. The NVIDIA card is the more powerful compute engine overall, but its utility in graphics-heavy, Metal-based applications is severely compromised.

The GTX 980 Ti’s nearest rivals, with average scores around 27,800-28,100, are the AMD Radeon Pro W5500X, AMD FirePro S7150, AMD Radeon RX 7800M, and AMD Radeon Pro Vega 20. These deltas are all within 0.7%, indicating the NVIDIA card sits in a crowded performance tier. The R9 M395X, similarly, is bracketed by the AMD FirePro W7100, AMD FirePro D700, and NVIDIA GeForce RTX 3080 Ti Mobile, all within 0.6% of its average score. This placement suggests the R9 M395X, despite its older architecture, still competes with much more modern mobile parts in aggregate. The GTX 980 Ti is the better choice for compute-heavy tasks and general OpenCL performance; the R9 M395X is the better choice for Metal-specific graphics workloads.

Architecture Differences

The architectural divide between these two GPUs is substantial. The NVIDIA GeForce GTX 980 Ti is built on the GM200 chip, utilizing the Maxwell 2.0 architecture from the GeForce 900 generation. It is fabricated on a 28 nm process at TSMC, with a massive 8,000 million transistors on a 601 mm² die, resulting in a transistor density of 13.3M / mm². In contrast, the AMD Radeon R9 M395X uses the Amethyst chip, based on the GCN 3.0 architecture from the R9 M300 generation. It also uses a 28 nm process at TSMC but integrates only 5,000 million transistors on a significantly smaller 366 mm² die, yielding a slightly higher density of 13.7M / mm².

Core configurations differ wildly. The NVIDIA card is equipped with 2,816 shading units, 176 texture mapping units (TMUs), and 96 render output units (ROPs). The AMD card has 2,048 shading units, 128 TMUs, and only 32 ROPs. This disparity in ROPs is a primary reason for the massive difference in pixel throughput: the GTX 980 Ti achieves 103.3 GPixel/s, while the R9 M395X manages only 23.14 GPixel/s. Similarly, texture rate is heavily in NVIDIA's favor, at 189.4 GTexel/s versus 92.54 GTexel/s.

Memory subsystems are also distinct. The GTX 980 Ti features 6 GB of GDDR5 memory on a 384-bit bus, delivering a bandwidth of 336.6 GB/s. The R9 M395X counters with a larger 8 GB frame buffer but on a narrower 256-bit bus, resulting in a much lower bandwidth of 160.0 GB/s. Clock speeds further separate them; the NVIDIA card has a base clock of 1000 MHz and a boost of 1076 MHz, while the AMD card's core clocks are not listed. The NVIDIA card's memory runs at 1753 MHz (7 Gbps effective), compared to the AMD's 1250 MHz (5 Gbps effective).

Power and physical design are polar opposites. The GTX 980 Ti has a TDP of 250 W, requires a 600 W power supply, uses a dual-slot cooler with 1x 6-pin and 1x 8-pin power connectors, and measures 267 mm in length. The R9 M395X is an MXM module with a TDP of just 75 W, no power connectors, and dimensions that are described as portable device dependent. In terms of API support, the NVIDIA card supports DirectX 12 (12_1) and Vulkan 1.4, while the AMD card supports DirectX 12 (12_0) and an older Vulkan 1.2.170. Both support OpenGL 4.6.

Head-to-Head Benchmarks

The head-to-head results reveal a stark binary outcome. There are only two shared benchmarks, and each GPU wins one decisively.

The first is Geekbench Metal, where the AMD Radeon R9 M395X wins with a score of 33,953. The NVIDIA GeForce GTX 980 Ti scores 19,520 in the same test. This represents a deltaPct of -42.5% for the NVIDIA card, meaning it is 42.5% slower than the AMD part in this specific workload. This is the single largest performance gap in the entire comparison. The AMD card's GCN architecture appears far more optimized for Metal API calls.

The second is Geekbench OpenCL, where the tables turn completely. The NVIDIA GeForce GTX 980 Ti scores 43,513, while the AMD Radeon R9 M395X scores just 17,829. This yields a deltaPct of 144.1% for the NVIDIA card, indicating it is 144.1% faster than the AMD card. This is an even larger margin than AMD's win in Metal. The raw compute power of the GTX 980 Ti, with its 6.060 TFLOPS of FP32 performance versus the AMD's 2.961 TFLOPS, is fully realized in the OpenCL environment. The AMD card does offer FP16 performance at a 1:1 ratio (2.961 TFLOPS), but this does not translate to an advantage in the OpenCL test.

These two results show a performance split that is almost perfectly inverse. The GTX 980 Ti's 144.1% lead in OpenCL is offset by the R9 M395X's 42.5% lead in Metal. While the overall benchmark average favors NVIDIA (28,020 vs 25,891), the real-world application of these scores is entirely dependent on which API the software uses.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 980 Ti has a higher average benchmark score of 28,020, compared to the AMD Radeon R9 M395X's 25,891.

Q: How much faster is the AMD Radeon R9 M395X in the Geekbench Metal test?

A: The AMD Radeon R9 M395X scores 33,953 in Geekbench Metal, making it 42.5% faster than the NVIDIA GeForce GTX 980 Ti's score of 19,520.

Q: What is the performance difference in the Geekbench OpenCL test?

A: In Geekbench OpenCL, the NVIDIA GeForce GTX 980 Ti scores 43,513, which is 144.1% higher than the AMD Radeon R9 M395X's score of 17,829.

Q: Do both GPUs have the same amount of memory?

A: No. The NVIDIA GeForce GTX 980 Ti has 6 GB of GDDR5 memory, while the AMD Radeon R9 M395X has 8 GB of GDDR5 memory.

Q: Which GPU has a higher transistor count?

A: The NVIDIA GeForce GTX 980 Ti has 8,000 million transistors, while the AMD Radeon R9 M395X has 5,000 million transistors.

Q: What is the TDP of each card?

A: The NVIDIA GeForce GTX 980 Ti has a TDP of 250 W, whereas the AMD Radeon R9 M395X has a much lower TDP of 75 W.

Where Each One Wins

The performance data dictates a clear use-case split based on API and workload type.

NVIDIA GeForce GTX 980 Ti wins in:

  • OpenCL Compute: The GTX 980 Ti is the undisputed winner here, with a 144.1% lead over the R9 M395X. Its FP32 performance of 6.060 TFLOPS is more than double the AMD card's 2.961 TFLOPS, making it vastly superior for general-purpose GPU compute tasks, rendering, and any application that relies on OpenCL.
  • Overall Aggregate Performance: With an average benchmark score of 28,020, it sits 8.2% higher than the R9 M395X's 25,891. This suggests it is a more powerful GPU in a broader sense, even if its specific API weaknesses hold it back.
  • Pixel and Texture Throughput: The GTX 980 Ti's 103.3 GPixel/s pixel rate and 189.4 GTexel/s texture rate are far superior to the R9 M395X's 23.14 GPixel/s and 92.54 GTexel/s, indicating a hardware advantage in rasterization-heavy tasks, provided the software uses a compatible API.

AMD Radeon R9 M395X wins in:

  • Metal Graphics: The R9 M395X is the definitive winner in the Geekbench Metal benchmark, scoring 33,953 to the GTX 980 Ti's 19,520. This 42.5% advantage makes it the superior choice for any application that uses Apple's Metal API, which is common in macOS environments.
  • Memory Capacity: The R9 M395X offers 8 GB of VRAM compared to the GTX 980 Ti's 6 GB. This larger frame buffer could be an advantage in high-resolution texture workloads, even if the bandwidth is significantly lower.
  • Power Efficiency: The R9 M395X's 75 W TDP is a fraction of the GTX 980 Ti's 250 W. This makes it the only viable option for portable devices and MXM module configurations, where power and cooling are constrained. The GTX 980 Ti's dual-slot, 267 mm design is entirely unsuitable for such platforms.

DETAILED SPECIFICATIONS

SPECIFICATION
R9 M395X
GTX 980 Ti
Core Specs
Shading Units
2,048
2,816 +37.5%
Shaders
2,048
2,816 +37.5%
TMUs
128
176 +37.5%
ROPs
32
96 +200.0%
Compute Units
32
—
Clocks
Base Clock
—
1000 MHz
Boost Clock
—
1076 MHz
GPU Clock
723 MHz
—
Memory Clock
1250 MHz 5 Gbps effective
1753 MHz 7 Gbps effective
Memory
Memory Size
8 GB
6 GB
VRAM (MB)
8,192
6,144 -25.0%
Memory Type
GDDR5
GDDR5
Memory Bus
256 bit
384 bit
Bandwidth
160.0 GB/s
336.6 GB/s
Cache
L1 Cache
16 KB (per CU)
48 KB (per SMM)
L2 Cache
512 KB
3 MB
Performance
Pixel Rate
23.14 GPixel/s
103.3 GPixel/s
Texture Rate
92.54 GTexel/s
189.4 GTexel/s
FP32 (TFLOPS)
2.961 TFLOPS
6.060 TFLOPS
FP64 (TFLOPS)
185.1 GFLOPS (1:16)
189.4 GFLOPS (1:32)
FP16 (TFLOPS)
2.961 TFLOPS (1:1)
—
Power
TDP
75 W
250 W
TDP (W)
75
250 +233.3%
Suggested PSU
—
600 W
Power Connectors
None
1x 6-pin + 1x 8-pin
Architecture
Architecture
GCN 3.0
Maxwell 2.0
GPU Name
Amethyst
GM200
Generation
Gem System (R9 M300)
GeForce 900
Process Size
28 nm
28 nm
Transistors
5,000 million
8,000 million
Die Size
366 mm²
601 mm²
Foundry
TSMC
TSMC
Density
13.7M / mm²
13.3M / mm²
API Support
DirectX
12 (12_0)
12 (12_1)
OpenGL
4.6
4.6
Vulkan
1.2.170
1.4
OpenCL
2.1
3.0
CUDA
—
5.2
Shader Model
6.5
6.8
Physical
Slot Width
MXM Module
Dual-slot
Length
—
267 mm 10.5 inches
Height
—
111 mm 4.4 inches
Outputs
Portable Device Dependent
1x DVI1x HDMI 2.03x DisplayPort 1.2
Bus Interface
PCIe 3.0 x16
PCIe 3.0 x16
Other
Launch Price
—
649 USD
Production
End-of-life
End-of-life
Predecessor
Solar System
GeForce 700
Successor
Polaris Mobile
GeForce 10
View Radeon R9 M395X Details View GeForce GTX 980 Ti Details