AMD FirePro S9300 X2 vs Intel Arc A370M Comparison

AMD
RADEON

AMD FirePro S9300 X2

CORE STATE Capsaicin
VRAM 4 GB
CLOCK SPEED
TDP 300 W
BUS WIDTH 4096 bit
ARCHITECTURE GCN 3.0
nm
PROCESS 28 nm
LAUNCH DATE 2016
VS
Intel
GPU

Arc A370M

CORE STATE DG2-128
VRAM 4 GB
CLOCK SPEED 2050 MHz
TDP 35 W
BUS WIDTH 64 bit
ARCHITECTURE Xe-HPG
nm
PROCESS 6 nm
LAUNCH DATE 2022

PERFORMANCE BENCHMARKS

geekbench_opencl
27,971
29,676
geekbench_vulkan
37,109
28,673

Analysis: AMD FirePro S9300 X2 vs Intel Arc A370M

Head-to-Head Benchmarks

The recorded data shows a split decision between these two GPUs, with each claiming one decisive victory in the two available benchmark tests. In Geekbench OpenCL, the Intel Arc A370M posts a score of 29676, which is 5.7% ahead of the AMD FirePro S9300 X2's 27971. This is a meaningful margin in a compute-oriented workload, and it flips the expected hierarchy given the FirePro's far larger silicon and power envelope.

The Vulkan results tell a very different story. The AMD FirePro S9300 X2 delivers 37109 points, a commanding 29.4% lead over the Intel Arc A370M's 28673. That is a substantial gap, nearly six times larger than Intel's OpenCL advantage in percentage terms. For any workload that scales well under Vulkan, the FirePro S9300 X2 is clearly the stronger part.

Looking at average benchmark scores across all recorded tests, the FirePro S9300 X2 lands at 32540, while the Arc A370M sits at 29175. That works out to roughly an 11.5% overall advantage for the AMD card. The FirePro also holds a higher percentile ranking among all GPUs, at 77 compared to Intel's 74. The nearest rivals for the FirePro S9300 X2 include the AMD Radeon RX 590 GME at 32601 (-0.2%), the AMD Radeon RX 7900 GRE at 32456 (+0.3%), the AMD FirePro S10000 at 32388 (+0.5%), and the NVIDIA T600 Mobile at 32849 (-0.9%). The Arc A370M's closest competitors are the AMD Radeon RX Vega M GH at 29197 (-0.1%), the AMD FirePro W8000 at 29211 (-0.1%), the AMD Radeon RX 470 at 28996 (+0.6%), and the AMD Radeon RX 6800M at 28874 (+1%).

The FirePro S9300 X2's average score places it in a tight pack where the spread between its nearest rivals is less than 1.5%, indicating that it is clustered with a group of similarly performing GPUs. The Arc A370M, meanwhile, sits in a slightly lower performance band, with its nearest rivals also clustered tightly around it. The 5.7% OpenCL deficit for the FirePro is notable, but the 29.4% Vulkan lead is the dominant data point in the head-to-head comparison.

Where Each One Wins

The Intel Arc A370M wins in OpenCL compute tasks. The 29676 score against 27971 shows a 5.7% advantage, and this holds across any OpenCL-based application: rendering, physics simulation, or general GPU compute that uses this API. For users whose primary workload is OpenCL, the Arc A370M is the better choice based on the benchmark data alone.

The AMD FirePro S9300 X2 wins decisively in Vulkan workloads. The 37109 score versus 28673 represents a 29.4% advantage, and this is the kind of gap that translates into clearly faster frame pacing in Vulkan-based games or Vulkan compute applications. The FirePro also holds the overall average benchmark lead, making it the stronger card when considering both recorded tests together.

The FirePro S9300 X2's higher average score of 32540 versus 29175 means that across the full test suite, it outperforms the Arc A370M in the aggregate. However, the Arc A370M's OpenCL win is not trivial; it suggests that API-specific optimization matters more than raw hardware capability in some cases. The FirePro has roughly twice the shading units (4096 versus 1024), four times the TMUs (256 versus 64), and double the ROPs (64 versus 32), yet still loses in OpenCL. This indicates that architecture efficiency and driver maturity in OpenCL favor the Intel part.

For users who care about a single dominant use case, the choice is clear: Vulkan users should pick the FirePro S9300 X2, while OpenCL users should pick the Arc A370M. For a balanced workload covering both APIs, the FirePro's aggregate lead makes it the safer pick.

Architecture Differences

The two GPUs come from completely different architectural lineages and manufacturing approaches. The AMD FirePro S9300 X2 uses the GCN 3.0 architecture, carrying the chip codename "Capsaicin" and belonging to the FirePro Server (Sx300) generation. It is built on a 28 nm process at TSMC, with 8,900 million transistors on a 596 mm² die, giving a transistor density of 14.9 million per square millimeter.

The Intel Arc A370M uses the Xe-HPG architecture, with the chip codename "DG2-128" and the generation label "Alchemist (Arc 3 Mobile)". It is built on a 6 nm process, also at TSMC, with 7,200 million transistors on a 157 mm² die, giving a transistor density of 45.9 million per square millimeter. The density difference is stark: Intel packs more than three times as many transistors per square millimeter despite having fewer total transistors.

Memory architecture differs substantially as well. The FirePro S9300 X2 uses 4 GB of HBM on a 4096-bit bus, yielding 512.0 GB/s of bandwidth. The Arc A370M uses 4 GB of GDDR6 on a 64-bit bus, yielding 112.0 GB/s. That is a 4.6x bandwidth advantage for the AMD card, which helps explain its Vulkan dominance. The FirePro's memory clock is listed at 500 MHz (1000 Mbps effective), while the Arc runs at 1750 MHz (14 Gbps effective), but the massive bus width difference overrides the clock speed advantage.

The FirePro S9300 X2 has no ray tracing cores, while the Arc A370M includes 8 ray tracing cores. Neither GPU lists tensor cores. The FirePro features 4096 shading units, 256 TMUs, and 64 ROPs. The Arc A370M has 1024 shading units, 64 TMUs, and 32 ROPs. The FirePro's pixel rate is 62.40 GPixel/s versus 65.60 GPixel/s for the Arc, a slight win for Intel despite the ROP deficit. The texture rate heavily favors AMD: 249.6 GTexel/s versus 131.2 GTexel/s. FP32 compute is 7.987 TFLOPS for the FirePro versus 4.198 TFLOPS for the Arc. The Arc does support FP16 at 8.397 TFLOPS (2:1 ratio), while no FP16 figure is recorded for the FirePro.

The FirePro S9300 X2 has no display outputs, making it a compute-only card. The Arc A370M has portable device dependent outputs, reflecting its mobile nature. The FirePro uses a PCIe 3.0 x16 interface, while the Arc uses PCIe 4.0 x8. API support differs: the FirePro supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The Arc supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The newer DirectX 12 Ultimate and Vulkan 1.4 support on the Intel card is notable, though the FirePro's Vulkan benchmark results show it still outperforms in that API.

FAQ

Q: Which GPU has higher average benchmark scores?

A: The AMD FirePro S9300 X2 has an average benchmark score of 32540, which is higher than the Intel Arc A370M's 29175.

Q: How large is the Vulkan performance gap?

A: The FirePro S9300 X2 scores 37109 in Geekbench Vulkan, which is 29.4% ahead of the Arc A370M's 28673.

Q: Does the Intel Arc A370M win any benchmark?

A: Yes, the Arc A370M wins the Geekbench OpenCL test with 29676 points, 5.7% ahead of the FirePro S9300 X2's 27971.

Q: What is the memory bandwidth difference?

A: The FirePro S9300 X2 has 512.0 GB/s of bandwidth from its 4096-bit HBM memory, while the Arc A370M has 112.0 GB/s from its 64-bit GDDR6 memory.

Q: Which GPU has ray tracing support?

A: The Intel Arc A370M has 8 ray tracing cores. The AMD FirePro S9300 X2 has no ray tracing cores recorded.

Q: What is the transistor density comparison?

A: The Intel Arc A370M has a transistor density of 45.9 million per square millimeter on its 6 nm process, while the AMD FirePro S9300 X2 has 14.9 million per square millimeter on its 28 nm process.

Specification Differences

| Specification | AMD FirePro S9300 X2 | Intel Arc A370M |

|---|---|---|

| Architecture | GCN 3.0 | Xe-HPG |

| Generation | FirePro Server (Sx300) | Alchemist (Arc 3 Mobile) |

| Process node | 28 nm | 6 nm |

| Transistors | 8,900 million | 7,200 million |

| Die size | 596 mm² | 157 mm² |

| Transistor density | 14.9M / mm² | 45.9M / mm² |

| Base clock | Not recorded | 1550 MHz |

| Boost clock | Not recorded | 2050 MHz |

| Memory type | HBM | GDDR6 |

| Memory bus width | 4096 bit | 64 bit |

| Memory bandwidth | 512.0 GB/s | 112.0 GB/s |

| Shading units | 4096 | 1024 |

| TMUs | 256 | 64 |

| ROPs | 64 | 32 |

| Ray tracing cores | Not recorded | 8 |

| Pixel rate | 62.40 GPixel/s | 65.60 GPixel/s |

| Texture rate | 249.6 GTexel/s | 131.2 GTexel/s |

| FP32 compute | 7.987 TFLOPS | 4.198 TFLOPS |

| FP16 compute | Not recorded | 8.397 TFLOPS (2:1) |

| TDP | 300 W | 35 W |

| Slot width | Dual-slot | IGP |

| Power connectors | 2x 8-pin | Not recorded |

| Suggested PSU | 700 W | Not recorded |

| Bus interface | PCIe 3.0 x16 | PCIe 4.0 x8 |

| Display outputs | No outputs | Portable Device Dependent |

| DirectX support | 12 (12_0) | 12 Ultimate (12_2) |

| Vulkan support | 1.2.170 | 1.4 |

| Dimensions | 267 mm length, 111 mm height | Not recorded |

| Release date | 2016-03-30 | 2022-03-29 |

| Launch MSRP | 5,999 USD | Not recorded |

The Verdict

The data presents two very different products that happen to share a 4 GB memory capacity and an end-of-life production status. The AMD FirePro S9300 X2 is a server-class compute card with no display outputs, a 300 W TDP, dual-slot cooling, and dual 8-pin power connectors. It demands a 700 W suggested PSU and occupies significant physical space at 267 mm in length. The Intel Arc A370M is a mobile IGP with a 35 W TDP, no separate power connectors, and no recorded dimensions, reflecting its integration into portable devices.

For Vulkan-centric workloads, the FirePro S9300 X2 is the clear pick. Its 29.4% lead in Geekbench Vulkan, combined with its 4.6x memory bandwidth advantage and nearly double the FP32 compute, makes it the stronger compute engine. The 7.987 TFLOPS FP32 figure versus 4.198 TFLOPS tells the raw capability story, and the 249.6 GTexel/s texture rate versus 131.2 GTexel/s reinforces it.

For OpenCL-specific tasks, the Arc A370M wins. The 5.7% lead in Geekbench OpenCL, despite having a quarter of the shading units and a fraction of the bandwidth, indicates that the Xe-HPG architecture and its drivers extract more efficiency from OpenCL workloads. The 6 nm process and much higher transistor density (45.9M / mm² versus 14.9M / mm²) likely contribute to this efficiency.

The overall average benchmark score favors the FirePro S9300 X2, and its higher percentile ranking (77 versus 74) confirms that it sits above the Arc in the broader GPU landscape. Users building a compute server with available power and space should choose the FirePro S9300 X2. Users needing a low-power mobile solution that handles OpenCL well should choose the Arc A370M. The FirePro's lack of display outputs makes it unsuitable for any workstation role requiring video output, while the Arc's portable device dependent outputs tie it to laptop implementations. The launch MSRP of 5,999 USD for the FirePro reflects its server positioning; the Arc A370M has no recorded launch MSRP.

DETAILED SPECIFICATIONS

SPECIFICATION
FirePro S9300 X2
A370M
Core Specs
Shading Units
4,096
1,024 -75.0%
Shaders
4,096
1,024 -75.0%
TMUs
256
64 -75.0%
ROPs
64
32 -50.0%
Compute Units
64
Execution Units
128
Clocks
Base Clock
1550 MHz
Boost Clock
2050 MHz
GPU Clock
975 MHz
Memory Clock
500 MHz 1000 Mbps effective
1750 MHz 14 Gbps effective
Memory
Memory Size
4 GB
4 GB
VRAM (MB)
4,096
4,096 0.0%
Memory Type
HBM
GDDR6
Memory Bus
4096 bit
64 bit
Bandwidth
512.0 GB/s
112.0 GB/s
Cache
L1 Cache
16 KB (per CU)
L2 Cache
2 MB
4 MB
Performance
Pixel Rate
62.40 GPixel/s
65.60 GPixel/s
Texture Rate
249.6 GTexel/s
131.2 GTexel/s
FP32 (TFLOPS)
7.987 TFLOPS
4.198 TFLOPS
FP64 (TFLOPS)
499.2 GFLOPS (1:16)
1,049.6 GFLOPS (1:4)
FP16 (TFLOPS)
8.397 TFLOPS (2:1)
AI/RT
RT Cores
8
XMX Cores
128
Power
TDP
300 W
35 W
TDP (W)
300
35 -88.3%
Suggested PSU
700 W
Power Connectors
2x 8-pin
Architecture
Architecture
GCN 3.0
Xe-HPG
GPU Name
Capsaicin
DG2-128
Generation
FirePro Server (Sx300)
Alchemist (Arc 3 Mobile)
Process Size
28 nm
6 nm
Transistors
8,900 million
7,200 million
Die Size
596 mm²
157 mm²
Foundry
TSMC
TSMC
Density
14.9M / mm²
45.9M / mm²
API Support
DirectX
12 (12_0)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.2.170
1.4
OpenCL
2.1
3.0
Shader Model
6.5
6.6
Physical
Slot Width
Dual-slot
IGP
Length
267 mm 10.5 inches
Height
111 mm 4.4 inches
Outputs
No outputs
Portable Device Dependent
Bus Interface
PCIe 3.0 x16
PCIe 4.0 x8
Other
Launch Price
5,999 USD
Production
End-of-life
End-of-life
Predecessor
FirePro Terascale
Successor
Radeon Pro GCN
View FirePro S9300 X2 Details View Arc A370M Details