Upload 24 files

.gitattributes

@@ -58,3 +58,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+benchmark_report.pdf filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,104 @@
 ---
 license: cc-by-4.0
+language:
+- en
+pretty_name: NVIDIA GB10 vs RTX 5090 Local LLM Benchmark
+tags:
+- benchmark
+- local-llm
+- qwen3-coder
+- nvidia
+- gb10
+- rtx-5090
+- lm-studio
+- tabular
 ---
+
+# NVIDIA GB10 vs. GeForce RTX 5090 - Local LLM Inference Benchmark
+
+**Model:** Qwen3-Coder-30B-A3B-Instruct  
+**Format:** GGUF, Q4_K_M, 18.63 GB  
+**Runtime:** LM Studio / llama.cpp  
+**Author:** Efehan A.  
+**Benchmark date:** 5 August 2026
+
+This repository contains a decode-focused local inference benchmark comparing an NVIDIA GB10 system with a Windows workstation containing two GeForce RTX 5090 GPUs. Telemetry shows that the inference workload was carried primarily by a single RTX 5090 (GPU 0), while GPU 1 remained mostly idle.
+
+## Headline result
+
+| Platform | Output | Runs | Mean tokens/s | Std. dev. | Mean TTFT | Mean generation time |
+|---|---:|---:|---:|---:|---:|---:|
+| NVIDIA GB10 | 512 | 5 | **44.93** | 0.02 | 0.0862 s | 11.374 s |
+| GeForce RTX 5090 | 512 | 5 | **243.46** | 1.32 | 0.0280 s | 2.099 s |
+| NVIDIA GB10 | 1024 | 5 | **43.80** | 0.03 | 0.1000 s | 23.355 s |
+| GeForce RTX 5090 | 1024 | 5 | **240.97** | 0.89 | 0.0320 s | 4.245 s |
+
+**Observed RTX 5090 throughput advantage:** 5.42x at 512 tokens and 5.50x at 1024 tokens.
+
+![Mean 512-token decode throughput](assets/chart_throughput_comparison.png)
+
+## Test configuration
+
+- Model: `Qwen3-Coder-30B-A3B-Instruct`
+- Architecture reported by LM Studio: `qwen3moe`
+- Quantization: `Q4_K_M`
+- GPU offload: 48 layers / maximum
+- Evaluation batch size: 512
+- Max concurrent predictions: 1
+- Unified KV cache: enabled
+- KV cache GPU offload: enabled
+- Flash Attention: enabled
+- K/V cache quantization: disabled
+- API request temperature: 0
+- One warm-up request followed by five measured requests per batch
+
+## Benchmark batches
+
+1. GB10, 512 completion tokens, five measured runs.
+2. GB10, 1024 completion tokens, five measured runs.
+3. RTX 5090, 512 completion tokens, five measured runs.
+4. RTX 5090, 1024 completion tokens, five measured runs.
+
+## Repeatability
+
+![Run stability](assets/chart_run_stability.png)
+
+Throughput coefficients of variation were approximately **0.04%** and **0.07%** on GB10, and **0.54%** and **0.37%** on RTX 5090. The cross-platform difference is much larger than run-to-run variation.
+
+## 1024-token sustained generation
+
+![1024-token throughput](assets/chart_output_scaling.png)
+
+At 1024 tokens, RTX 5090 averaged **240.97 tokens/s** versus **43.80 tokens/s** on GB10. Doubling output length reduced throughput by only **2.50%** on GB10 and **1.02%** on RTX 5090.
+
+## Interpretation
+
+The tested 18.63 GB model fits inside the RTX 5090's 32 GB GDDR7 memory. NVIDIA specifies 1,792 GB/s peak memory bandwidth for the RTX 5090, compared with 273 GB/s for the GB10 unified-memory platform. The measured decode gap is directionally consistent with the large bandwidth difference, although operating system, runtime version and kernel differences prevent single-cause attribution.
+
+The systems target different deployment priorities:
+
+- **RTX 5090:** maximum single-stream decode speed when the model fits in 32 GB.
+- **GB10:** 128 GB unified-memory capacity for larger models, larger KV caches and longer-context workloads that do not fit on a single RTX 5090.
+
+## Important limitations
+
+- The 512-token benchmark output reports context 8192 on GB10 and 10132 on RTX, despite load-status captures showing 8192 on both systems. The corrected 1024-token results report context 8192 on both systems.
+- Runtime builds differed: ARM64/Linux CUDA 13 versus x86-64/Windows CUDA 12.
+- CPU thread settings differed: 20 on GB10 and 8 on RTX.
+- No direct prompt-prefill benchmark was completed.
+- Power efficiency is not compared because GB10 SoC telemetry and discrete-GPU board power are not directly equivalent.
+- Only one model, quantization and inference application were tested.
+
+## Files
+
+- [`benchmark_report.pdf`](benchmark_report.pdf) - full professional report
+- [`benchmark_summary.csv`](benchmark_summary.csv) - aggregate results
+- [`run_level_results.csv`](run_level_results.csv) - all measured runs
+- [`evidence/`](evidence/) - public-release screenshots
+
+## References
+
+1. NVIDIA DGX Spark Hardware Overview: https://docs.nvidia.com/dgx/dgx-spark/hardware.html
+2. NVIDIA GeForce RTX 5090 Specifications: https://www.nvidia.com/en-us/geforce/graphics-cards/50-series/rtx-5090/
+3. NVIDIA RTX Blackwell GPU Architecture: https://images.nvidia.com/aem-dam/Solutions/geforce/blackwell/nvidia-rtx-blackwell-gpu-architecture.pdf
+4. LM Studio REST API v0: https://lmstudio.ai/docs/developer/rest/endpoints

benchmark_report.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6410f323a9fc73d185f3e92a66115148aa329834cd5dda1b9cf13ee1a0ea52c
+size 2131581

benchmark_summary.csv

@@ -0,0 +1,5 @@
+Platform,Output tokens / batch,Runs,TPS mean,TPS stdev,TTFT mean s,TTFT stdev s,Generation mean s,Generation stdev s,Reported context,Quant
+NVIDIA GB10,512,5,44.926,0.018165902124586224,0.0862,0.0209093280618962,11.3742,0.004549725266431083,8192,Q4_K_M
+GeForce RTX 5090,512,5,243.458,1.3222594299153276,0.028,0.009354143466934854,2.099,0.011423659658795946,10132,Q4_K_M
+NVIDIA GB10,1024,5,43.802,0.029495762407507453,0.09999999999999999,0.040218155104380406,23.3552,0.01386362146049874,8192,Q4_K_M
+GeForce RTX 5090,1024,5,240.96800000000002,0.8928997704109899,0.032,0.010954451150103323,4.2454,0.015725775020646888,8192,Q4_K_M

run_level_results.csv

@@ -0,0 +1,21 @@
+platform,batch,run,completion_tokens,tokens_per_second,ttft_seconds,generation_seconds,reported_context,quant
+NVIDIA GB10,512-token,1,512,44.9,0.107,11.381,8192,Q4_K_M
+NVIDIA GB10,512-token,2,512,44.92,0.111,11.376,8192,Q4_K_M
+NVIDIA GB10,512-token,3,512,44.93,0.071,11.373,8192,Q4_K_M
+NVIDIA GB10,512-token,4,512,44.95,0.069,11.369,8192,Q4_K_M
+NVIDIA GB10,512-token,5,512,44.93,0.073,11.372,8192,Q4_K_M
+NVIDIA GB10,1024-token,1,1024,43.81,0.146,23.351,8192,Q4_K_M
+NVIDIA GB10,1024-token,2,1024,43.79,0.142,23.36,8192,Q4_K_M
+NVIDIA GB10,1024-token,3,1024,43.81,0.073,23.353,8192,Q4_K_M
+NVIDIA GB10,1024-token,4,1024,43.76,0.07,23.375,8192,Q4_K_M
+NVIDIA GB10,1024-token,5,1024,43.84,0.069,23.337,8192,Q4_K_M
+GeForce RTX 5090,512-token,1,512,244.85,0.042,2.087,10132,Q4_K_M
+GeForce RTX 5090,512-token,2,512,243.8,0.033,2.096,10132,Q4_K_M
+GeForce RTX 5090,512-token,3,512,242.07,0.02,2.111,10132,Q4_K_M
+GeForce RTX 5090,512-token,4,512,244.5,0.024,2.09,10132,Q4_K_M
+GeForce RTX 5090,512-token,5,512,242.07,0.021,2.111,10132,Q4_K_M
+GeForce RTX 5090,1024-token,1,1024,242.24,0.043,4.223,8192,Q4_K_M
+GeForce RTX 5090,1024-token,2,1024,240.31,0.041,4.257,8192,Q4_K_M
+GeForce RTX 5090,1024-token,3,1024,241.5,0.035,4.236,8192,Q4_K_M
+GeForce RTX 5090,1024-token,4,1024,240.08,0.022,4.261,8192,Q4_K_M
+GeForce RTX 5090,1024-token,5,1024,240.71,0.019,4.25,8192,Q4_K_M