NVIDIA L40G
NVIDIA graphics card specifications and benchmark scores
Ray Tracing
Tensor CoresAt a Glance
NVIDIANVIDIA L40G Specifications
GPU Core
Shader units and compute resources
The NVIDIA L40G GPU core specifications define its raw processing power for graphics and compute workloads. Shading units (also called CUDA cores, stream processors, or execution units depending on manufacturer) handle the parallel calculations required for rendering. TMUs (Texture Mapping Units) process texture data, while ROPs (Render Output Units) handle final pixel output. Higher shader counts generally translate to better GPU benchmark performance, especially in demanding games and 3D applications.
L40G Clock Speeds
GPU and memory frequencies
Clock speeds directly impact the L40G's performance in GPU benchmarks and real-world gaming. The base clock represents the minimum guaranteed frequency, while the boost clock indicates peak performance under optimal thermal conditions. Memory clock speed affects texture loading and frame buffer operations. The L40G by NVIDIA dynamically adjusts frequencies based on workload, temperature, and power limits to maximize performance while maintaining stability.
NVIDIA's L40G Memory
VRAM capacity and bandwidth
VRAM (Video RAM) is dedicated memory for storing textures, frame buffers, and shader data. The L40G's memory capacity determines how well it handles high-resolution textures and multiple displays. Memory bandwidth, measured in GB/s, affects how quickly data moves between the GPU and VRAM. Higher bandwidth improves performance in memory-intensive scenarios like 4K gaming. The memory bus width and type (GDDR6, GDDR6X, HBM) significantly influence overall GPU benchmark scores.
L40G by NVIDIA Cache
On-chip cache hierarchy
On-chip cache provides ultra-fast data access for the L40G, reducing the need to fetch data from slower VRAM. L1 and L2 caches store frequently accessed data close to the compute units. AMD's Infinity Cache (L3) dramatically increases effective bandwidth, improving GPU benchmark performance without requiring wider memory buses. Larger cache sizes help maintain high frame rates in memory-bound scenarios and reduce power consumption by minimizing VRAM accesses.
L40G Theoretical Performance
Compute and fill rates
Theoretical performance metrics provide a baseline for comparing the NVIDIA L40G against other graphics cards. FP32 (single-precision) performance, measured in TFLOPS, indicates compute capability for gaming and general GPU workloads. FP64 (double-precision) matters for scientific computing. Pixel and texture fill rates determine how quickly the GPU can render complex scenes. While real-world GPU benchmark results depend on many factors, these specifications help predict relative performance levels.
L40G Ray Tracing & AI
Hardware acceleration features
The NVIDIA L40G includes dedicated hardware for ray tracing and AI acceleration. RT cores handle real-time ray tracing calculations for realistic lighting, reflections, and shadows in supported games. Tensor cores (NVIDIA) or XMX cores (Intel) accelerate AI workloads including DLSS, FSR, and XeSS upscaling technologies. These features enable higher visual quality without proportional performance costs, making the L40G capable of delivering both stunning graphics and smooth frame rates in modern titles.
Ada Lovelace Architecture & Process
Manufacturing and design details
The NVIDIA L40G is built on NVIDIA's Ada Lovelace architecture, which defines how the GPU processes graphics and compute workloads. The manufacturing process node affects power efficiency, thermal characteristics, and maximum clock speeds. Smaller process nodes pack more transistors into the same die area, enabling higher performance per watt. Understanding the architecture helps predict how the L40G will perform in GPU benchmarks compared to previous generations.
Power & Thermal
TDP and power requirements
Power specifications for the NVIDIA L40G determine PSU requirements and thermal management needs. TDP (Thermal Design Power) indicates the heat output under typical loads, guiding cooler selection. Power connector requirements ensure adequate power delivery for stable operation during demanding GPU benchmarks. The suggested PSU wattage accounts for the entire system, not just the graphics card. Efficient power delivery enables the L40G to maintain boost clocks without throttling.
L40G by NVIDIA Physical & Connectivity
Dimensions and outputs
Physical dimensions of the NVIDIA L40G are critical for case compatibility. Card length, height, and slot width determine whether it fits in your chassis. The PCIe interface version affects bandwidth for communication with the CPU. Display outputs define monitor connectivity options, with modern cards supporting multiple high-resolution displays simultaneously. Verify these specifications against your case and motherboard before purchasing to ensure a proper fit.
NVIDIA API Support
Graphics and compute APIs
API support determines which games and applications can fully utilize the NVIDIA L40G. DirectX 12 Ultimate enables advanced features like ray tracing and variable rate shading. Vulkan provides cross-platform graphics capabilities with low-level hardware access. OpenGL remains important for professional applications and older games. CUDA (NVIDIA) and OpenCL enable GPU compute for video editing, 3D rendering, and scientific applications. Higher API versions unlock newer graphical features in GPU benchmarks and games.
L40G Product Information
Release and pricing details
The NVIDIA L40G is manufactured by NVIDIA as part of their graphics card lineup. Release date and launch pricing provide context for comparing GPU benchmark results with competing products from the same era. Understanding the product lifecycle helps evaluate whether the L40G by NVIDIA represents good value at current market prices. Predecessor and successor information aids in tracking generational improvements and planning future upgrades.
About NVIDIA L40G
NVIDIA L40G is an end-of-life server accelerator built on the Ada Lovelace architecture, using the AD102 chip fabricated on TSMC’s 5 nm process. The GPU packs 76,300 million transistors on a 609 mm² die, yielding a transistor density of 125.3M per mm². It is positioned as a server-grade compute card with a dual-slot form factor, PCIe 4.0 x16 interface, and a 300 W TDP, placing it in the 50th percentile of all GPUs in the database with an average benchmark score of zero. The following analysis interprets the available specification data, focusing on performance, positioning, ray tracing, power, and practical considerations.
Benchmark Performance
The L40G’s raw compute metrics are substantial, though the benchmark database lists no actual workload scores. The FP32 throughput reaches 89.97 TFLOPS, which is identical to its FP16 rating at 89.97 TFLOPS (1:1 ratio). This 1:1 ratio indicates the card does not rely on reduced-precision acceleration for FP16 workloads, a design choice favoring consistent throughput across mixed-precision tasks. The texture rate stands at 1,405.8 GTexel/s, while the pixel rate is 475.2 GPixel/s. These figures derive from 18,176 shading units, 568 texture mapping units, and 192 raster output units.
Memory bandwidth is 864.0 GB/s, delivered via 24 GB of GDDR6 on a 384-bit bus, with an effective data rate of 18 Gbps. For comparison, the nearestRivals array is empty, so no direct percentage deltas against specific competitors can be cited. However, the 50th percentile ranking suggests the L40G sits at the median of all GPUs in the database, implying that while its absolute numbers are high, other server or workstation cards likely exceed it in certain metrics. The lack of benchmark scores means the percentile is derived from specification-based heuristics rather than measured results. In practical terms, the FP32 figure of 89.97 TFLOPS positions the card for heavy parallel compute, but without rival data, relative performance cannot be quantified beyond that percentile.
How It Compares
The FACT PACK lists no nearest rivals, so direct head-to-head comparisons with named products are impossible. The predecessor is “Server Ampere” and the successor is “Server Hopper,” which indicates generational placement: the L40G bridges the previous Ampere server line and the subsequent Hopper architecture. Because no rival scores or deltaPct values are provided, the analysis must rely on architectural context. The Ada Lovelace generation is a newer design than Ampere, typically offering improved ray tracing and tensor core efficiency per watt, but the L40G’s 300 W TDP and 24 GB memory capacity are moderate for a server card. The 50th percentile ranking suggests it is neither a top-tier nor entry-level part in the broader GPU landscape. Without specific rival data, any claim about being “ahead” or “behind” a particular product would be unsupported. The only factual comparison is generational: it succeeds Server Ampere and precedes Server Hopper, meaning it inherits Ada’s feature set while lacking Hopper’s newer innovations.
Ray Tracing and Feature Set
The L40G includes 142 ray tracing cores and 568 tensor cores, both integral to Ada Lovelace’s feature set. Ray tracing cores accelerate BVH traversal and ray-triangle intersection, while tensor cores handle matrix operations for AI inference and DLSS-type workloads. The API support is comprehensive: DirectX 12 Ultimate (feature level 12_2), OpenGL 4.6, and Vulkan 1.4. DirectX 12 Ultimate implies support for hardware ray tracing, variable rate shading, and mesh shaders, all standard for modern GPU architectures. Vulkan 1.4 is a recent API revision, indicating long-term driver compatibility for cross-platform compute. The display outputs include 1x HDMI 2.1 and 3x DisplayPort 1.4a, which is unusual for a server card but suggests possible use in visualization or edge rendering scenarios. The FP16 and FP32 both run at 89.97 TFLOPS, meaning the tensor cores do not provide a separate FP16 boost, which is typical for some Ada server SKUs that prioritize consistent precision. The 142 RT cores and 568 tensor cores are specific counts from the FACT PACK; no comparative ratios to other cards are available. The pixel rate of 475.2 GPixel/s and texture rate of 1,405.8 GTexel/s indicate ample rasterization throughput, though server workloads likely emphasize compute over rendering.
FAQ
Q: What is the memory configuration of the NVIDIA L40G?
A: The L40G has 24 GB of GDDR6 memory on a 384-bit bus, with a bandwidth of 864.0 GB/s and an effective data rate of 18 Gbps.
Q: Does the L40G support hardware ray tracing?
A: Yes, it includes 142 ray tracing cores, and its API support includes DirectX 12 Ultimate (12_2), which mandates hardware ray tracing capabilities.
Q: What is the FP32 compute performance?
A: The FP32 throughput is 89.97 TFLOPS, and the FP16 performance is identical at 89.97 TFLOPS with a 1:1 ratio.
Q: What power connector and PSU are required?
A: The card uses a single 16-pin power connector, and the suggested power supply is 700 W. The TDP is rated at 300 W.
Q: Is the L40G still in production?
A: No, the production status is marked as end-of-life. Its release date was 2022-10-12, with the predecessor being Server Ampere and the successor being Server Hopper.
Q: What is the physical size of the card?
A: The L40G is a dual-slot card with a length of 267 mm (10.5 inches) and a height of 111 mm (4.4 inches). The width is not specified.
Power and Cooling
The L40G has a thermal design power of 300 W, which is moderate for a server accelerator. The suggested power supply is 700 W, indicating that the card’s peak transient loads are expected to stay within that envelope when combined with a typical CPU and other components. The power delivery uses a single 16-pin connector, which is the modern PCIe 5.0-style connector, though the card itself runs on PCIe 4.0 x16. The dual-slot form factor implies a passive or active cooling solution sufficient for the 300 W TDP; the exact cooler type is not specified in the FACT PACK, but the dual-slot width allows for a substantial heatsink. The card’s dimensions—267 mm long and 111 mm high—fit standard server chassis, with no width provided. Given the 300 W TDP, a capable air cooler or server airflow is expected to manage thermals, but no specific cooling hardware is named. The end-of-life status suggests that power efficiency may be less competitive than newer Hopper parts, but the 300 W figure is a hard spec from the FACT PACK. For system integration, the 700 W PSU recommendation is the only guidance provided; no alternative power configurations are listed. The single 16-pin connector simplifies cabling, but users must ensure their PSU has that connector or an adapter, which is not detailed in the data.
Detailed benchmark scores and charts for the NVIDIA L40G are below.
Benchmark Scores
No benchmark data available for this GPU.
Compare with Other GPUs
Select another GPU to compare specifications and benchmarks side-by-side.
Browse GPUs