Intel Arc A380E vs NVIDIA Jetson T5000 Comparison

The Verdict

The database records two fundamentally different hardware products: the Intel Arc A380E, an end-of-life desktop graphics card from the Alchemist generation, and the NVIDIA Jetson T5000, an active embedded/server module based on the Blackwell architecture. The recorded specifications show they serve distinct purposes, with the Arc A380E targeting display-centric workloads and the Jetson T5000 focusing on compute and AI acceleration. The Arc A380E carries a 75 W TDP and 4.096 TFLOPS of FP32 performance, while the Jetson T5000 operates at 120 W and delivers 8.064 TFLOPS FP32, meaning the NVIDIA part roughly doubles the raw single-precision throughput. The Jetson T5000 also provides 128 GB of LPDDR5X memory versus 6 GB of GDDR6 on the Intel card, a 20-fold capacity advantage that points toward large-model inference or data-parallel workloads. However, the Arc A380E includes 4x DisplayPort 2.0 outputs and full DirectX 12 Ultimate support, while the Jetson T5000 has no display outputs and lists N/A for all common graphics APIs, so it is not a replacement for a traditional rendering GPU. Buyers needing a video output and gaming-capable graphics should select the Arc A380E; users requiring high-bandwidth memory and tensor-core compute in a compact module should select the Jetson T5000.

FAQ

Q: Which product has higher FP32 performance?

A: The NVIDIA Jetson T5000 delivers 8.064 TFLOPS, exactly double the 4.096 TFLOPS of the Intel Arc A380E.

Q: How much memory does each product offer?

A: The Intel Arc A380E has 6 GB of GDDR6 on a 96-bit bus, while the NVIDIA Jetson T5000 has 128 GB of LPDDR5X on a 256-bit bus.

Q: Does the Jetson T5000 support DirectX?

A: No, the database lists DirectX as N/A for the Jetson T5000, whereas the Arc A380E supports DirectX 12 Ultimate (12_2).

Q: What is the thermal design power difference?

A: The Arc A380E is rated at 75 W, and the Jetson T5000 is rated at 120 W, a 45 W difference.

Q: Which product has display outputs?

A: Only the Intel Arc A380E has display outputs, specifically 4x DisplayPort 2.0; the Jetson T5000 has none.

Q: What are the tensor-core counts?

A: The Jetson T5000 includes 96 tensor cores; the Arc A380E has no listed tensor cores.

Architecture Differences

The Intel Arc A380E uses the Xe-HPG architecture on the DG2-128 chip, belonging to the Alchemist (Arc 3) generation. It is fabricated on a 6 nm process at TSMC, with 7,200 million transistors on a 157 mm² die, yielding a transistor density of 45.9M per mm². The chip integrates 1,024 shading units, 64 texture mapping units, 32 ROPs, and 8 ray-tracing cores. FP16 performance reaches 8.192 TFLOPS at a 2:1 ratio relative to FP32, reflecting a consumer-oriented design where half-precision throughput is doubled via specialized execution. The card supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, making it compatible with standard graphics stacks and gaming APIs.

The NVIDIA Jetson T5000 uses the Blackwell architecture on the GB10B chip, from the Server Blackwell (Bxx) generation. It is built on a 5 nm process at TSMC, with an unknown transistor count and a die size of 391 mm². The module houses 2,560 shading units, 80 TMUs, 32 ROPs, 20 ray-tracing cores, and 96 tensor cores. FP16 performance is 8.064 TFLOPS at a 1:1 ratio, meaning the Jetson T5000 does not gain extra throughput for half-precision operations, a design choice oriented toward consistent compute across precisions. The architecture targets server or embedded deployments, as evidenced by the absence of display outputs and the N/A entries for DirectX, OpenGL, and Vulkan. The tensor core count of 96 is a clear indicator of AI-focused hardware, whereas the Arc A380E lists no tensor cores at all.

The process node difference is small but notable: 6 nm for Intel versus 5 nm for NVIDIA. The die size gap is substantial, with the Jetson T5000 measuring 391 mm² versus 157 mm² for the Arc A380E, even though the transistor count for the Jetson T5000 is not recorded. The memory subsystem also diverges sharply, with the Arc A380E using GDDR6 and the Jetson T5000 using LPDDR5X, each paired with different bus widths and capacities.

Specification Differences

The two products differ across nearly every major specification field in the database. Clock speeds: the Arc A380E runs at a fixed 2000 MHz base and boost, while the Jetson T5000 has a 1386 MHz base and 1575 MHz boost. Memory clocks follow a similar pattern, with the Arc A380E at 1937 MHz (15.5 Gbps effective) and the Jetson T5000 at 1067 MHz (8.5 Gbps effective). Memory capacity and type are the largest divergences: 6 GB GDDR6 on a 96-bit bus for Intel versus 128 GB LPDDR5X on a 256-bit bus for NVIDIA. Resulting bandwidth is 186.0 GB/s for the Arc A380E and 273.2 GB/s for the Jetson T5000, a 47% advantage for the latter.

Compute resources differ in count and type. The Arc A380E has 1,024 shading units, 64 TMUs, 32 ROPs, and 8 RT cores, with no tensor cores. The Jetson T5000 has 2,560 shading units, 80 TMUs, 32 ROPs, 20 RT cores, and 96 tensor cores. Pixel rates are 64.00 GPixel/s for Intel versus 50.40 GPixel/s for NVIDIA, while texture rates are 128.0 GTexel/s versus 126.0 GTexel/s, a near tie. FP32 output favors NVIDIA at 8.064 TFLOPS versus 4.096 TFLOPS, and FP16 output favors Intel at 8.192 TFLOPS versus 8.064 TFLOPS, though the Intel figure relies on a 2:1 ratio.

Power and physical design: the Arc A380E draws 75 W and fits a single slot with no power connectors, while the Jetson T5000 draws 120 W and is an IGP (integrated graphics processor) module with no power connectors. The suggested PSU is 250 W for Intel and 300 W for NVIDIA. Bus interfaces differ, with PCIe 4.0 x8 on the Arc A380E and PCIe 5.0 x8 on the Jetson T5000. Dimensions are starkly different: the Arc A380E measures 254 mm by 127 mm by 20 mm, while the Jetson T5000 measures 87 mm by 100 mm by 15 mm, making the NVIDIA module far more compact. The Arc A380E has 4x DisplayPort 2.0 outputs and supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4; the Jetson T5000 has no outputs and lists N/A for all three APIs. Production status shows the Arc A380E as end-of-life, while the Jetson T5000 is active. The release dates are March 31, 2024, for Intel and August 26, 2025, for NVIDIA. The launch MSRP for the Jetson T5000 is 2,999 USD, stated once here, while the Arc A380E has no recorded launch MSRP.

Head-to-Head Benchmarks

The database contains no recorded head-to-head benchmark scores for these two products, and each has zero wins in the winsA and winsB fields. Both products sit at the 50th percentile versus all GPUs, with an average benchmark score of zero. This absence of measured results means direct performance comparisons must rely on the specification-derived metrics rather than application-level tests.

The FP32 compute gap is the clearest differentiator. The Jetson T5000 delivers 8.064 TFLOPS, exactly twice the 4.096 TFLOPS of the Arc A380E. This doubling appears in shading unit counts as well, with the Jetson T5000 housing 2,560 shading units versus 1,024 for the Arc A380E. However, the Arc A380E posts a higher pixel rate at 64.00 GPixel/s versus 50.40 GPixel/s, a 27% advantage that likely stems from its higher boost clock (2000 MHz versus 1575 MHz) and its ROP count, which matches at 32. Texture rates are effectively equal at 128.0 GTexel/s for Intel and 126.0 GTexel/s for NVIDIA, a difference of only 1.6%.

In FP16, the Arc A380E edges ahead with 8.192 TFLOPS versus 8.064 TFLOPS, a 1.6% lead, but the Intel figure uses a 2:1 ratio that the NVIDIA part does not employ. The Jetson T5000 maintains 1:1 FP16 performance, which means its FP16 throughput equals its FP32 throughput, a characteristic that may appeal to mixed-precision workloads.

Memory bandwidth favors the Jetson T5000 at 273.2 GB/s versus 186.0 GB/s, a 46.9% advantage. The 128 GB capacity dwarfs the 6 GB of the Arc A380E, and the 256-bit bus doubles the 96-bit bus of the Intel card. These memory figures suggest the Jetson T5000 can hold larger datasets and move them faster, though the Arc A380E compensates with higher effective memory clock speed (15.5 Gbps versus 8.5 Gbps).

Ray-tracing resources also favor the Jetson T5000 with 20 RT cores versus 8, a 2.5x difference, and the tensor core count of 96 versus none gives NVIDIA a clear lead in AI-accelerated tasks. The Arc A380E counters with full graphics API support, including DirectX 12 Ultimate, while the Jetson T5000 has none.

Where Each One Wins

The Intel Arc A380E wins in scenarios that require standard graphics output and rendering. Its 4x DisplayPort 2.0 connectors enable multi-monitor setups, and its support for DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 makes it suitable for gaming, content creation, or any workload that relies on conventional GPU APIs. The higher pixel rate of 64.00 GPixel/s and the 2000 MHz fixed clock support rasterization-heavy tasks, and the single-slot design with no power connectors simplifies installation into standard desktop systems. The 75 W TDP also means it can run in systems with a modest 250 W suggested PSU, and the end-of-life status may appeal to buyers seeking a legacy or replacement part for existing Alchemist-based systems. The FP16 advantage of 8.192 TFLOPS, achieved through the 2:1 ratio, gives it a slight edge in half-precision compute if the workload can use that path.

The NVIDIA Jetson T5000 wins in compute-dense and AI-focused deployments. Its 96 tensor cores and 8.064 TFLOPS FP32 performance position it for neural network inference, training, or scientific computing where FP32 throughput matters. The 128 GB LPDDR5X memory with 273.2 GB/s bandwidth allows large model weights or datasets to reside on-chip, avoiding frequent transfers over the PCIe 5.0 x8 interface. The compact dimensions of 87 mm by 100 mm by 15 mm and the IGP form factor make it appropriate for embedded systems, edge servers, or dense chassis where space is limited. The 20 RT cores add ray-tracing capability, though the lack of display outputs and standard graphics APIs means it will not drive a monitor directly. The active production status and 5 nm process indicate a current-generation part, and the 120 W TDP with a 300 W suggested PSU fits a power envelope suitable for server environments.

The Arc A380E also wins on texture rate by a marginal 128.0 GTexel/s versus 126.0 GTexel/s, which could matter in texturing-bound rendering pipelines, but the Jetson T5000 dominates in shading units (2,560 versus 1,024), tensor cores (96 versus none), memory capacity (128 GB versus 6 GB), and FP32 throughput (8.064 TFLOPS versus 4.096 TFLOPS). The choice reduces to whether the workload demands display output and graphics API compatibility, which favors the Intel card, or high-capacity memory and tensor-core compute, which favors the NVIDIA module.