Intel Graphics 24EU Mobile vs NVIDIA N1X 48SM Comparison

Where Each One Wins

The benchmark data in this comparison shows a fundamental split in workload suitability between the Intel Graphics 24EU Mobile and the NVIDIA N1X 48SM. Each part is optimized for a different class of tasks, and the recorded specifications confirm that these are not direct competitors in most scenarios.

The Intel Graphics 24EU Mobile is built for low-power, integrated graphics duty inside a compact system. It uses the Xe-LP architecture on a 10 nm Intel process, with a base clock of 300 MHz and a boost clock of 1000 MHz. Its shading units number 192, with 12 texture mapping units and 4 raster output units. The pixel rate is 4.000 GPixel/s and the texture rate is 12.00 GTexel/s. Floating point performance reaches 384.0 GFLOPS in FP32 and 768.0 GFLOPS in FP16 with a 2:1 ratio. The thermal design power is 6 W, which places it firmly in the segment of ultra-low-power graphics silicon intended for portable devices.

The NVIDIA N1X 48SM, by contrast, is a substantially larger and more capable part. It uses the Blackwell 2.0 architecture on a 5 nm TSMC process, with a die size of 382 mm². The base clock is 741 MHz and the boost clock is 2346 MHz. It carries 6144 shading units, 384 texture mapping units, 48 raster output units, 48 ray tracing cores, and 192 tensor cores. Pixel rate is 112.6 GPixel/s, texture rate is 900.9 GTexel/s, and FP32 performance is 28.83 TFLOPS. FP16 performance is also 28.83 TFLOPS with a 1:1 ratio. Memory consists of 128 GB of LPDDR5X on a 256 bit bus, delivering 273.2 GB/s of bandwidth.

The wins are defined by which part can execute a given workload at a meaningful performance level. The Intel part wins in scenarios where power envelope and system integration matter most, such as basic display output, lightweight 2D acceleration, and low-overhead media playback. The NVIDIA part wins in every compute-heavy scenario: high-resolution rendering, ray tracing workloads, tensor core acceleration, and large memory buffer operations. The data shows that the Intel Graphics 24EU Mobile has no ray tracing cores and no tensor cores, while the NVIDIA N1X 48SM has 48 and 192 respectively. That alone determines the outcome for any workload that relies on those fixed-function units.

Architecture Differences

The two parts come from different architectural families and different foundries. Intel uses the Xe-LP architecture, manufactured on its own 10 nm process. NVIDIA uses Blackwell 2.0, manufactured by TSMC on a 5 nm process. The die size for the NVIDIA part is 382 mm², while the Intel die size is not recorded in the database. Transistor counts are unknown for both parts.

The Intel Graphics 24EU Mobile belongs to the HD Graphics-T generation, specifically the Twin Lake chip. It is an integrated graphics processor with a Ring Bus interface. The NVIDIA N1X 48SM belongs to the Blackwell IGP generation, using the GB20B chip, and connects via PCIe 5.0 x16.

Memory architecture differs sharply. The Intel part uses system shared memory, with a system dependent bandwidth figure and no dedicated memory bus width. The NVIDIA part uses 128 GB of LPDDR5X with a 256 bit bus and a fixed bandwidth of 273.2 GB/s. Memory clock for NVIDIA is 1067 MHz, described as 8.5 Gbps effective. The Intel memory clock is listed as "System Shared" with no dedicated frequency.

Compute resources diverge in every category. The Intel part has 192 shading units, 12 TMUs, and 4 ROPs. The NVIDIA part has 6144 shading units, 384 TMUs, and 48 ROPs. The NVIDIA part includes 48 ray tracing cores and 192 tensor cores; the Intel part has none. FP32 throughput is 384.0 GFLOPS for Intel versus 28.83 TFLOPS for NVIDIA, a difference of two orders of magnitude. FP16 throughput is 768.0 GFLOPS for Intel versus 28.83 TFLOPS for NVIDIA, with Intel using a 2:1 ratio and NVIDIA using a 1:1 ratio.

API support also differs. The Intel part supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The NVIDIA part lists DirectX as N/A, OpenGL as N/A, and Vulkan as N/A. This means the Intel part has a defined API surface for graphics workloads, while the NVIDIA part's API support is not recorded in the database, which suggests it may not be intended for conventional graphics API workloads in the same manner.

The display outputs are also different. Intel lists "Portable Device Dependent" outputs, while NVIDIA lists a single HDMI output. The Intel part has no power connectors, consistent with its IGP slot width and 6 W TDP. The NVIDIA part also has no power connectors, also an IGP, but its TDP is unknown. The NVIDIA part uses PCIe 5.0 x16 as its bus interface, which is a much wider and faster connection than the Ring Bus used by Intel.

FAQ

Q: Which part has a higher boost clock?

A: The NVIDIA N1X 48SM has a boost clock of 2346 MHz, while the Intel Graphics 24EU Mobile boosts to 1000 MHz. The NVIDIA part also has a higher base clock at 741 MHz versus 300 MHz for Intel.

Q: Does either part support ray tracing?

A: The NVIDIA N1X 48SM includes 48 ray tracing cores. The Intel Graphics 24EU Mobile has no ray tracing cores, so ray tracing workloads are only possible on the NVIDIA part.

Q: How much memory does each part use?

A: The NVIDIA N1X 48SM has 128 GB of LPDDR5X on a 256 bit bus with 273.2 GB/s bandwidth. The Intel Graphics 24EU Mobile uses system shared memory with system dependent bandwidth.

Q: What is the process node for each chip?

A: The Intel Graphics 24EU Mobile uses a 10 nm process from Intel. The NVIDIA N1X 48SM uses a 5 nm process from TSMC.

Q: Which part has tensor cores?

A: The NVIDIA N1X 48SM has 192 tensor cores. The Intel Graphics 24EU Mobile has no tensor cores.

Q: What are the FP32 performance figures?

A: The Intel Graphics 24EU Mobile delivers 384.0 GFLOPS in FP32. The NVIDIA N1X 48SM delivers 28.83 TFLOPS in FP32.

Specification Differences

The database records the following differences between the two parts:

• Manufacturer: Intel versus NVIDIA
• Chip: Twin Lake versus GB20B
• Architecture: Xe-LP versus Blackwell 2.0
• Generation: HD Graphics-T (Twin Lake) versus Blackwell IGP (N1x)
• Process node: 10 nm versus 5 nm
• Foundry: Intel versus TSMC
• Die size: unknown versus 382 mm²
• Base clock: 300 MHz versus 741 MHz
• Boost clock: 1000 MHz versus 2346 MHz
• Memory clock: System Shared versus 1067 MHz 8.5 Gbps effective
• Memory size: System Shared versus 128 GB
• Memory type: System Shared versus LPDDR5X
• Memory bus width: System Shared versus 256 bit
• Memory bandwidth: System Dependent versus 273.2 GB/s
• Shading units: 192 versus 6144
• TMUs: 12 versus 384
• ROPs: 4 versus 48
• RT cores: none versus 48
• Tensor cores: none versus 192
• Pixel rate: 4.000 GPixel/s versus 112.6 GPixel/s
• Texture rate: 12.00 GTexel/s versus 900.9 GTexel/s
• FP32: 384.0 GFLOPS versus 28.83 TFLOPS
• FP16: 768.0 GFLOPS (2:1) versus 28.83 TFLOPS (1:1)
• TDP: 6 W versus unknown
• Bus interface: Ring Bus versus PCIe 5.0 x16
• Display outputs: Portable Device Dependent versus 1x HDMI
• DirectX: 12 (12_1) versus N/A
• OpenGL: 4.6 versus N/A
• Vulkan: 1.4 versus N/A
• Release date: 2024-12-31 versus 2026-05-31

The only shared fields are slot width (IGP for both), power connectors (none for both), production status (Active for both), and the absence of a launch MSRP in the database for either part.

Head-to-Head Benchmarks

The recorded data shows no direct benchmark scores for either part, and the head-to-head benchmark list is empty. The nearest rivals lists are also empty, meaning the database has no comparative third-party scores for these two parts. However, the specification data allows for a direct comparison of theoretical throughput and feature capability.

The largest win for the NVIDIA N1X 48SM is in FP32 compute. The Intel part delivers 384.0 GFLOPS, while the NVIDIA part delivers 28.83 TFLOPS. Expressed as a ratio, the NVIDIA part is roughly 75 times faster in raw FP32 throughput. This is the dominant metric for general purpose compute workloads, and the gap is so large that no software optimization on the Intel side could close it.

In FP16, the NVIDIA part also wins decisively. Intel delivers 768.0 GFLOPS with a 2:1 ratio, meaning it halves its FP32 rate to achieve FP16. NVIDIA delivers 28.83 TFLOPS with a 1:1 ratio, meaning it sustains its full FP32 rate in FP16. The NVIDIA part is approximately 37.5 times faster in FP16 throughput.

Texture rate favors NVIDIA by a similar margin. The Intel part achieves 12.00 GTexel/s, while the NVIDIA part achieves 900.9 GTexel/s. This is a 75x advantage, consistent with the TMU count difference of 12 versus 384. Pixel rate also favors NVIDIA, with 112.6 GPixel/s versus 4.000 GPixel/s for Intel, a 28x advantage driven by the ROP count difference of 48 versus 4.

Memory bandwidth is another decisive win for NVIDIA. The Intel part has system dependent bandwidth, which in practice is limited by the host platform's memory subsystem. The NVIDIA part has a fixed 273.2 GB/s from its dedicated LPDDR5X memory on a 256 bit bus. For workloads that are memory bound, such as large buffer operations or high-resolution texture streaming, the NVIDIA part has a structural advantage that does not depend on the host system.

The ray tracing and tensor core comparison is binary. Intel has none, NVIDIA has 48 ray tracing cores and 192 tensor cores. Any workload that uses ray tracing, such as realistic lighting simulation, can only run on the NVIDIA part. Any workload that uses tensor cores, such as matrix multiplication for machine learning inference, can also only run on the NVIDIA part.

The Intel part does have a defined API surface with DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The NVIDIA part lists N/A for all three APIs, which means the database does not record conventional graphics API support for it. This is a notable difference in the opposite direction: for traditional graphics API workloads, the Intel part has documented support while the NVIDIA part does not.

Power consumption is a win for Intel, with a 6 W TDP recorded. The NVIDIA part has an unknown TDP, so no direct comparison is possible, but the scale of the compute resources on the NVIDIA part suggests it would require substantially more power to sustain those clocks across 6144 shading units.

The release dates differ by approximately 17 months, with Intel launching at the end of 2024 and NVIDIA launching in mid-2026. Both parts are marked as Active in production status. The NVIDIA part has a die size of 382 mm², while the Intel die size is unknown. The NVIDIA part uses PCIe 5.0 x16, while Intel uses a Ring Bus interface. These differences in bus interface and die size further confirm that the NVIDIA part is a much larger, more complex piece of silicon.

The data indicates that the Intel Graphics 24EU Mobile is designed for scenarios where power draw is the limiting factor and graphics workloads are light. The NVIDIA N1X 48SM is designed for scenarios where throughput is the limiting factor and power is less constrained. The two parts do not meaningfully overlap in their performance envelopes.