Intel UHD Graphics 770 Mobile vs NVIDIA N1 20SM Comparison

The Verdict

The recorded data places these two integrated GPUs in different performance classes despite both occupying the IGP slot form factor. The Intel UHD Graphics 770 Mobile, built on Raptor Lake with Generation 12.2 architecture, delivers 819.2 GFLOPS of FP32 compute, while the NVIDIA N1 20SM, based on the GB20B chip with Blackwell 2.0 architecture, delivers 12.01 TFLOPS of FP32 compute. That represents a 14.7x gap in raw floating-point throughput. The N1 20SM also carries 20 RT cores and 80 tensor cores, features absent entirely from the Intel part. Benchmark results indicate the NVIDIA solution is the stronger option for any workload that can use its compute resources, ray tracing, or tensor acceleration. The Intel part remains viable for basic display output and lightweight tasks, but the data does not support choosing it for performance-sensitive applications.

Where Each One Wins

The Intel UHD Graphics 770 Mobile wins in compatibility and integration simplicity. It uses system shared memory, has no dedicated power connectors, and runs on a Ring Bus interface. Its DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4 API support means it can run modern graphics APIs on Windows-based systems. The texture rate of 25.60 GTexel/s and pixel rate of 12.80 GPixel/s are sufficient for 2D workloads, video playback, and undemanding legacy titles. The 15 W TDP figure places it in a low-power envelope suitable for thin portable devices.

The NVIDIA N1 20SM wins in every measurable compute category. Its 2560 shading units compare to 256 on the Intel part, a 10x difference. The 160 texture mapping units versus 16 represent a 10x advantage. The 24 ROPs versus 8 give it a 3x edge in pixel output. Memory bandwidth shows a massive divide: 273.2 GB/s from a 256-bit LPDDR5X interface versus system dependent bandwidth on the Intel side. The 128 GB memory capacity, while unusual for a GPU, indicates the N1 20SM targets a different class of device entirely. The 20 RT cores and 80 tensor cores open up hardware-accelerated ray tracing and AI inference, neither of which the Intel part can offer.

Architecture Differences

The two chips come from different foundries and process nodes. Intel uses its own 10 nm process for the Raptor Lake chip, while NVIDIA uses TSMC's 5 nm process for the GB20B chip. The Intel architecture is Generation 12.2, which is the HD Graphics-M lineage from the Raptor Lake generation. The NVIDIA architecture is Blackwell 2.0, from the Blackwell IGP (N1x) generation. The foundry difference matters: TSMC's 5 nm node allows higher transistor density and clock speeds, contributing to the N1 20SM's 2346 MHz boost clock versus 1600 MHz on the Intel part.

The NVIDIA die size is listed at 382 mm², while the Intel die size is not recorded in the database. NVIDIA's transistor count is marked unknown, and Intel's is not recorded either. The memory subsystem differs fundamentally: Intel uses system shared memory with system dependent bandwidth, while NVIDIA has a dedicated 128 GB LPDDR5X pool on a 256-bit bus. The NVIDIA memory clock runs at 1067 MHz with 8.5 Gbps effective transfer rate. The NVIDIA chip also includes dedicated RT and tensor hardware, which the Intel chip lacks entirely. The Intel part has no RT cores and no tensor cores, while NVIDIA provides 20 and 80 respectively.

The bus interface differs: Intel uses Ring Bus, NVIDIA uses PCIe 5.0 x16. Display outputs also differ: Intel lists "Portable Device Dependent" while NVIDIA provides a single HDMI output. The NVIDIA part has no power connectors, and its TDP is unknown in the database. The Intel part has a 15 W TDP. API support shows a stark contrast: Intel supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, while NVIDIA lists N/A for DirectX, OpenGL, and Vulkan. This suggests the N1 20SM targets a non-standard software stack, possibly Linux-based or custom environments, rather than traditional Windows gaming.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The NVIDIA N1 20SM delivers 12.01 TFLOPS of FP32 compute, while the Intel UHD Graphics 770 Mobile delivers 819.2 GFLOPS. The NVIDIA part is roughly 14.7x faster in this metric.

Q: Does the Intel UHD Graphics 770 Mobile support hardware ray tracing?

A: No. The Intel part has no RT cores listed in the database. The NVIDIA N1 20SM includes 20 RT cores.

Q: What memory configuration does each GPU use?

A: The Intel UHD Graphics 770 Mobile uses system shared memory with system dependent bandwidth. The NVIDIA N1 20SM uses 128 GB of LPDDR5X memory on a 256-bit bus with 273.2 GB/s bandwidth.

Q: Which GPU supports DirectX 12?

A: The Intel UHD Graphics 770 Mobile supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The NVIDIA N1 20SM lists N/A for DirectX, OpenGL, and Vulkan.

Q: What are the clock speeds for each GPU?

A: The Intel part runs at a 300 MHz base clock and 1600 MHz boost clock. The NVIDIA part runs at 741 MHz base and 2346 MHz boost.

Q: Which GPU has more shading units?

A: The NVIDIA N1 20SM has 2560 shading units, compared to 256 on the Intel UHD Graphics 770 Mobile, a 10x difference.

Head-to-Head Benchmarks

The database shows no direct head-to-head benchmark scores for these two parts, so the comparison relies on recorded specifications that translate directly into compute throughput. The most significant single metric is FP32 performance. The NVIDIA N1 20SM produces 12.01 TFLOPS versus 819.2 GFLOPS on the Intel part. This 14.7x gap in raw FP32 throughput means any shader-heavy workload will complete dramatically faster on the NVIDIA solution.

Texture throughput follows a similar pattern. The NVIDIA part reaches 375.4 GTexel/s against 25.60 GTexel/s on Intel, a 14.7x advantage. This ratio matches the FP32 gap because texture rate scales with shading unit count and clock speed. The Intel part has 16 TMUs at up to 1600 MHz, while NVIDIA has 160 TMUs at up to 2346 MHz. Pixel rate shows a smaller but still substantial gap: 56.30 GPixel/s versus 12.80 GPixel/s, a 4.4x difference. The NVIDIA part has 24 ROPs versus 8 on Intel, and the higher clock contributes further.

Memory bandwidth is where the NVIDIA part pulls away most decisively. The N1 20SM has a dedicated 256-bit LPDDR5X interface delivering 273.2 GB/s, while the Intel part relies on system shared memory with bandwidth listed as system dependent. In practice, shared system memory typically provides far lower bandwidth than dedicated VRAM, so the real-world gap is likely larger than the nominal numbers suggest. The NVIDIA memory clock of 1067 MHz with 8.5 Gbps effective transfer rate supports this high bandwidth figure.

FP16 performance shows a different relationship between the two parts. The Intel UHD Graphics 770 Mobile delivers 1.638 TFLOPS with a 2:1 ratio to FP32, meaning it halves throughput for FP16. The NVIDIA N1 20SM delivers 12.01 TFLOPS with a 1:1 ratio, meaning it sustains full throughput for FP16. This gives NVIDIA a 7.3x advantage in FP16 workloads, a smaller gap than the FP32 comparison but still decisive. The 1:1 FP16 ratio on NVIDIA indicates dedicated FP16 hardware, while the 2:1 ratio on Intel suggests the FP16 path shares resources with FP32.

The RT core and tensor core counts add capabilities that the Intel part cannot match at any clock speed. The NVIDIA N1 20SM includes 20 RT cores for ray tracing and 80 tensor cores for AI acceleration. The Intel part has neither. For workloads that use these units, the NVIDIA part is not just faster, it is the only option. The shading unit count difference, 2560 versus 256, means even without RT or tensor acceleration, the NVIDIA part has 10x the parallel execution units.

The clock speed difference also favors NVIDIA. The boost clock of 2346 MHz on the N1 20SM compares to 1600 MHz on the Intel part. Higher clocks amplify the already substantial core count advantage. The base clocks show a different relationship: NVIDIA at 741 MHz versus Intel at 300 MHz, but the boost clocks matter more for sustained performance. The Intel part maintains a 15 W TDP, while the NVIDIA TDP is unknown, so the power efficiency comparison cannot be made from the recorded data.

Specification Differences

The two GPUs differ in nearly every recorded specification. The process node shows Intel at 10 nm versus NVIDIA at 5 nm. The foundry is Intel for the Intel part and TSMC for the NVIDIA part. The chip names differ: Raptor Lake versus GB20B. The architectures differ: Generation 12.2 versus Blackwell 2.0. The generation names differ: HD Graphics-M (Raptor Lake) versus Blackwell IGP (N1x).

Clock specifications show Intel at 300 MHz base and 1600 MHz boost, while NVIDIA runs at 741 MHz base and 2346 MHz boost. Memory clocks differ: Intel lists "System Shared" while NVIDIA lists 1067 MHz with 8.5 Gbps effective. Memory size: Intel uses system shared, NVIDIA has 128 GB. Memory type: Intel uses system shared, NVIDIA uses LPDDR5X. Bus width: Intel uses system shared, NVIDIA uses 256 bit. Bandwidth: Intel lists system dependent, NVIDIA lists 273.2 GB/s.

Compute unit counts differ across the board. Shading units: 256 versus 2560. TMUs: 16 versus 160. ROPs: 8 versus 24. RT cores: absent versus 20. Tensor cores: absent versus 80. Pixel rate: 12.80 GPixel/s versus 56.30 GPixel/s. Texture rate: 25.60 GTexel/s versus 375.4 GTexel/s. FP32: 819.2 GFLOPS versus 12.01 TFLOPS. FP16: 1.638 TFLOPS (2:1) versus 12.01 TFLOPS (1:1).

Power and interface details differ. TDP: 15 W on Intel, unknown on NVIDIA. Slot width is IGP for both. Power connectors: absent on NVIDIA, not listed on Intel. Bus interface: Ring Bus on Intel versus PCIe 5.0 x16 on NVIDIA. Display outputs: "Portable Device Dependent" on Intel versus 1x HDMI on NVIDIA.

API support shows the widest gap. Intel supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. NVIDIA lists N/A for all three. Production status is Active for both. Release dates differ: Intel launched on 2023-01-03, NVIDIA on 2026-05-31. The Intel part has a successor listed as Arc Graphics-M, while the NVIDIA part has no successor. Die size is recorded only for NVIDIA at 382 mm². Transistor count is unknown for NVIDIA and not recorded for Intel. Neither part has a launch MSRP, benchmark scores, or nearest rivals in the database. Both sit at the 50th percentile against all GPUs with an average benchmark score of 0.