RX 7900 XTX image benchmarks
589 reported configurations · 5 sources · 625 measurements
Documented results
24 selected configurationsSelected across sources, not a performance ranking. Match model weights, resolution, steps, software and timing scope before comparing.
| Source | Model | GPUs | Output | Measurements |
|---|---|---|---|---|
| View ReportUL Procyon | Stable Diffusion 1.5Text → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 4100 stepsBatch 4 | 3.1 s/imageTime / image · reported Whole workflow · Timing details missing |
| View ReportDiffusion workflow | Stable Diffusion XLText → image | Radeon RX 7900 XTX1 GPUConfiguration
| Resolution not reportedSteps not reportedBatch not reported | 6.3 s/imageTime / image · reported Whole workflow · Timing details missing |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 1024 × 1024Batch 18 stepsBatch 1 | 31.493 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 27.594 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 1.603 svae decode time s · reportedVAE component, not whole-image time. 22,740 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View Reportstable-diffusion.cpp / sd2R | FLUX.2-klein-4BText → image | Radeon RX 7900 XTXGPU count not reportedConfiguration
| 1024 × 10248 stepsBatch not reported | 9.56 sReported time · reportedSource-reported duration; the arm does not explicitly restate the timer boundary. Timing details missing |
| View ReportSD WebUI | juggernautxl_juggxibyrundiffusionText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 1Steps not reportedBatch 1 | 6.34 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |
| View ReportUL Procyon | Stable Diffusion XLText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 1024 × 1024Batch 1100 stepsBatch 1 | 25.2 s/imageTime / image · reported Whole workflow · Timing details missing |
| View ReportDiffusion workflow | Stable Diffusion XLText → image | Radeon RX 7900 XTX1 GPUConfiguration
| Resolution not reportedSteps not reportedBatch not reported | 0.292 s/forwardCalculated · Model forward time · reported · Model componentExact ms-to-seconds conversion; isolated UNet, not image latency.reported milliseconds / 1000 Model component · Timing details missing |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 18 stepsBatch 1 | 10.225 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 7.797 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 0.334 svae decode time s · reportedVAE component, not whole-image time. 22,260 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View Reportstable-diffusion.cpp / sd2R | FLUX.2-klein-4BText → image | Radeon RX 7900 XTXGPU count not reportedConfiguration
| 1024 × 10248 stepsBatch not reported | 11.9 sSampling time · reportedExplicit sampling time. Sampling only · Timing details missing |
| View ReportSD WebUI | juggernautxl_juggxibyrundiffusionText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 2Steps not reportedBatch 2 | 8.15 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 576 × 1024Batch 18 stepsBatch 1 | 17.719 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 14.441 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 0.761 svae decode time s · reportedVAE component, not whole-image time. 22,452 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View Reportstable-diffusion.cpp / sd2R | FLUX.2-klein-4BText → image | Radeon RX 7900 XTXGPU count not reportedConfiguration
| 1024 × 10248 stepsBatch not reported | 12.3 sReported time · reportedSource-reported duration; the arm does not explicitly restate the timer boundary. Timing details missing |
| View ReportSD WebUI | juggernautxl_juggxibyrundiffusionText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 4Steps not reportedBatch 4 | 7.43 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 1024 × 1024Batch 18 stepsBatch 1 | 23.463 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 21.771 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 1.618 svae decode time s · reportedVAE component, not whole-image time. 16,436 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View Reportstable-diffusion.cpp / sd2R | FLUX.2-klein-4BText → image | Radeon RX 7900 XTXGPU count not reportedConfiguration
| 1024 × 10248 stepsBatch not reported | 12.4 sReported time · reportedSource-reported duration; the arm does not explicitly restate the timer boundary. Timing details missing |
| View ReportSD WebUI | mixillustriousmerge1_mixillustriousmerge3Text → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 1Steps not reportedBatch 1 | 6.6 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 18 stepsBatch 1 | 5.676 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 5.286 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 0.313 svae decode time s · reportedVAE component, not whole-image time. 15,940 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View Reportstable-diffusion.cpp / sd2R | FLUX.2-klein-4BText → image | Radeon RX 7900 XTXGPU count not reportedConfiguration
| 1024 × 10248 stepsBatch not reported | 13.9 sReported time · reportedSource-reported duration; the arm does not explicitly restate the timer boundary. Timing details missing |
| View ReportSD WebUI | mixillustriousmerge1_mixillustriousmerge3Text → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 2Steps not reportedBatch 2 | 1.7 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 576 × 1024Batch 18 stepsBatch 1 | 11.913 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 11.064 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 0.775 svae decode time s · reportedVAE component, not whole-image time. 16,132 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View Reportstable-diffusion.cpp / sd2R | Z-Image TurboText → image | Radeon RX 7900 XTXGPU count not reportedConfiguration
| 1024 × 10248 stepsBatch not reported | 14.38 sReported time · reportedSource-reported duration; the arm does not explicitly restate the timer boundary. Timing details missing |
| View ReportSD WebUI | mixillustriousmerge1_mixillustriousmerge3Text → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 4Steps not reportedBatch 4 | 1.79 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |
| View ReportDiffusion workflow | Z-Image TurboText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 1024 × 1024Batch 18 stepsBatch 1 | 23.273 sReported time · reportedFull GPU graph wall clock; model loading, tokenization, graph setup and file writing excluded. 21.581 sSampling time · reportedDiT component profiler time inside the full graph; do not add to graph duration. 1.616 svae decode time s · reportedVAE component, not whole-image time. 16,166 MiBmemory footprint mib · reportedSource definition: free VRAM before minus free VRAM after workspace; not a sampled transient peak. Model component · Timing details missing |
| View ReportSD WebUI | sdxl1.0\cattowernoobaixl_v15vpred\cattowernoobaixl_v15vpred.safetensorsText → image | Radeon RX 7900 XTX1 GPUConfiguration
| 512 × 512Batch 1Steps not reportedBatch 1 | 7.44 image-steps/sPipeline throughput · reportedHarness computes steps times batch divided by process_images wall time; not sampler iterations per second or images per second. |