Upload performance_data.md
Browse files- performance_data.md +134 -0
performance_data.md
ADDED
|
@@ -0,0 +1,134 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Markdown# Performance Benchmarks and Test Results
|
| 2 |
+
|
| 3 |
+
**Invention Title:** Method for Ternary-Quantized Transformer Optimization with Buffered Routing Embedding and SWA Attention
|
| 4 |
+
|
| 5 |
+
**Inventor:** Konstantin Vladimirovich Grabko
|
| 6 |
+
**Test Date:** December 2025
|
| 7 |
+
**Hardware Tested:** AMD MI50 (32 GB HBM2), custom cooling
|
| 8 |
+
|
| 9 |
+
**Confidentiality Notice:** Internal test data – proprietary and not for publication.
|
| 10 |
+
-
|
| 11 |
+
- JiRackPyTorch_BitNet_class_3b.py
|
| 12 |
+
-
|
| 13 |
+
# ROCm System Management Interface
|
| 14 |
+
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
## Concise Info
|
| 18 |
+
|
| 19 |
+
| GPU | Temp (DieEdge) | AvgPwr | SCLK | MCLK | Fan | Perf | PwrCap | VRAM% | GPU% |
|
| 20 |
+
|-------|----------------|--------|------------|-----------|--------|--------|---------|-------|------|
|
| 21 |
+
| GPU[0] | 46.0c | N/A | 1725Mhz | 1000Mhz | 17.65% | auto | 225.0W | 59% | 100% |
|
| 22 |
+
|
| 23 |
+
- **GPU[0]:** `get_power_avg` is not supported on the given system.
|
| 24 |
+
|
| 25 |
+
---
|
| 26 |
+
|
| 27 |
+
### End of ROCm SMI Log
|
| 28 |
+
-
|
| 29 |
+
- Step 4270 | Loss: 9.1875 | VRAM: 14.85GB | 11.9 t/s
|
| 30 |
+
- Step 4275 | Loss: 9.0000 | VRAM: 14.84GB | 11.9 t/s
|
| 31 |
+
- Step 4280 | Loss: 9.3125 | VRAM: 14.84GB | 11.9 t/s
|
| 32 |
+
- Step 4285 | Loss: 10.6875 | VRAM: 14.85GB | 11.9 t/s
|
| 33 |
+
- Step 4290 | Loss: 10.1250 | VRAM: 14.84GB | 11.9 t/s
|
| 34 |
+
- Step 4295 | Loss: 10.4375 | VRAM: 14.84GB | 11.9 t/s
|
| 35 |
+
- Step 4300 | Loss: 10.6250 | VRAM: 14.84GB | 11.9 t/s
|
| 36 |
+
- Step 4305 | Loss: 10.3125 | VRAM: 14.84GB | 11.9 t/s
|
| 37 |
+
- Step 4310 | Loss: 10.3750 | VRAM: 14.84GB | 11.9 t/s
|
| 38 |
+
- Step 4315 | Loss: 10.8750 | VRAM: 14.85GB | 11.9 t/s
|
| 39 |
+
- Step 4320 | Loss: 10.1875 | VRAM: 14.84GB | 11.9 t/s
|
| 40 |
+
- Step 4325 | Loss: 9.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 41 |
+
- Step 4330 | Loss: 9.7500 | VRAM: 14.84GB | 11.9 t/s
|
| 42 |
+
- Step 4335 | Loss: 8.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 43 |
+
- Step 4340 | Loss: 9.1875 | VRAM: 14.85GB | 11.9 t/s
|
| 44 |
+
- Step 4345 | Loss: 10.3125 | VRAM: 14.84GB | 11.9 t/s
|
| 45 |
+
- Step 4350 | Loss: 10.9375 | VRAM: 14.84GB | 11.9 t/s
|
| 46 |
+
- Step 4355 | Loss: 10.5000 | VRAM: 14.84GB | 11.9 t/s
|
| 47 |
+
- Step 4360 | Loss: 9.1250 | VRAM: 14.84GB | 11.9 t/s
|
| 48 |
+
- Step 4365 | Loss: 10.4375 | VRAM: 14.84GB | 11.9 t/s
|
| 49 |
+
- Step 4370 | Loss: 9.3125 | VRAM: 14.84GB | 11.9 t/s
|
| 50 |
+
- Step 4375 | Loss: 10.0625 | VRAM: 14.84GB | 11.9 t/s
|
| 51 |
+
- Step 4380 | Loss: 10.3125 | VRAM: 14.85GB | 11.9 t/s
|
| 52 |
+
- Step 4385 | Loss: 10.5000 | VRAM: 14.84GB | 11.9 t/s
|
| 53 |
+
- Step 4390 | Loss: 9.5000 | VRAM: 14.84GB | 11.9 t/s
|
| 54 |
+
- Step 4395 | Loss: 10.9375 | VRAM: 14.85GB | 11.9 t/s
|
| 55 |
+
- Step 4400 | Loss: 7.5312 | VRAM: 14.85GB | 11.9 t/s
|
| 56 |
+
- Step 4405 | Loss: 9.7500 | VRAM: 14.84GB | 11.9 t/s
|
| 57 |
+
- Step 4410 | Loss: 10.7500 | VRAM: 14.85GB | 11.9 t/s
|
| 58 |
+
- Step 4415 | Loss: 9.1875 | VRAM: 14.84GB | 11.9 t/s
|
| 59 |
+
- Step 4420 | Loss: 11.0000 | VRAM: 14.84GB | 11.9 t/s
|
| 60 |
+
- Step 4425 | Loss: 9.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 61 |
+
- Step 4430 | Loss: 10.3750 | VRAM: 14.84GB | 11.9 t/s
|
| 62 |
+
- Step 4435 | Loss: 10.8750 | VRAM: 14.84GB | 11.9 t/s
|
| 63 |
+
- Step 4440 | Loss: 10.9375 | VRAM: 14.85GB | 11.9 t/s
|
| 64 |
+
- Step 4445 | Loss: 10.0000 | VRAM: 14.84GB | 11.9 t/s
|
| 65 |
+
- Step 4450 | Loss: 9.1875 | VRAM: 14.84GB | 11.9 t/s
|
| 66 |
+
- Step 4455 | Loss: 9.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 67 |
+
- Step 4460 | Loss: 10.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 68 |
+
- Step 4465 | Loss: 10.4375 | VRAM: 14.85GB | 11.9 t/s
|
| 69 |
+
- Step 4470 | Loss: 10.4375 | VRAM: 14.84GB | 11.9 t/s
|
| 70 |
+
- Step 4475 | Loss: 9.5000 | VRAM: 14.84GB | 11.9 t/s
|
| 71 |
+
- Step 4480 | Loss: 9.8750 | VRAM: 14.85GB | 11.9 t/s
|
| 72 |
+
- Step 4485 | Loss: 8.1875 | VRAM: 14.85GB | 11.9 t/s
|
| 73 |
+
- Step 4490 | Loss: 11.1875 | VRAM: 14.84GB | 11.9 t/s
|
| 74 |
+
- Step 4495 | Loss: 10.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 75 |
+
- Step 4500 | Loss: 10.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 76 |
+
>>> SAVING: Checkpoint to ./models/ternary_3b_checkpoint_- Step_4500...
|
| 77 |
+
>>> CLEANUP: Removing old checkpoint ./models/ternary_3b_checkpoint_- Step_3000
|
| 78 |
+
- Step 4505 | Loss: 10.5000 | VRAM: 14.85GB | 11.9 t/s
|
| 79 |
+
- Step 4510 | Loss: 9.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 80 |
+
- Step 4515 | Loss: 9.8750 | VRAM: 14.84GB | 11.9 t/s
|
| 81 |
+
- Step 4520 | Loss: 9.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 82 |
+
- Step 4525 | Loss: 10.6250 | VRAM: 14.84GB | 11.9 t/s
|
| 83 |
+
- Step 4530 | Loss: 9.3750 | VRAM: 14.85GB | 11.9 t/s
|
| 84 |
+
- Step 4535 | Loss: 9.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 85 |
+
- Step 4540 | Loss: 10.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 86 |
+
- Step 4545 | Loss: 11.0000 | VRAM: 14.84GB | 11.9 t/s
|
| 87 |
+
- Step 4550 | Loss: 10.0000 | VRAM: 14.85GB | 11.9 t/s
|
| 88 |
+
- Step 4555 | Loss: 9.9375 | VRAM: 14.84GB | 11.9 t/s
|
| 89 |
+
- Step 4560 | Loss: 11.0625 | VRAM: 14.85GB | 11.9 t/s
|
| 90 |
+
- Step 4565 | Loss: 9.3125 | VRAM: 14.85GB | 11.9 t/s
|
| 91 |
+
- Step 4570 | Loss: 9.3750 | VRAM: 14.84GB | 11.9 t/s
|
| 92 |
+
- Step 4575 | Loss: 10.8125 | VRAM: 14.85GB | 11.9 t/s
|
| 93 |
+
- Step 4580 | Loss: 10.7500 | VRAM: 14.85GB | 11.9 t/s
|
| 94 |
+
- Step 4585 | Loss: 9.3750 | VRAM: 14.85GB | 11.9 t/s
|
| 95 |
+
- Step 4590 | Loss: 10.7500 | VRAM: 14.84GB | 11.9 t/s
|
| 96 |
+
- Step 4595 | Loss: 9.3125 | VRAM: 14.84GB | 11.9 t/s
|
| 97 |
+
- Step 4600 | Loss: 10.6250 | VRAM: 14.84GB | 11.9 t/s
|
| 98 |
+
- Step 4605 | Loss: 10.4375 | VRAM: 14.84GB | 11.9 t/s
|
| 99 |
+
- Step 4610 | Loss: 9.8750 | VRAM: 14.85GB | 11.9 t/s
|
| 100 |
+
- Step 4615 | Loss: 10.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 101 |
+
- Step 4620 | Loss: 10.0625 | VRAM: 14.85GB | 11.9 t/s
|
| 102 |
+
- Step 4625 | Loss: 10.6250 | VRAM: 14.84GB | 11.9 t/s
|
| 103 |
+
- Step 4630 | Loss: 10.7500 | VRAM: 14.85GB | 11.9 t/s
|
| 104 |
+
- Step 4635 | Loss: 10.5000 | VRAM: 14.84GB | 11.9 t/s
|
| 105 |
+
- Step 4640 | Loss: 10.0000 | VRAM: 14.85GB | 11.9 t/s
|
| 106 |
+
- Step 4645 | Loss: 10.9375 | VRAM: 14.84GB | 11.9 t/s
|
| 107 |
+
- Step 4650 | Loss: 10.6250 | VRAM: 14.84GB | 11.9 t/s
|
| 108 |
+
- Step 4655 | Loss: 9.6875 | VRAM: 14.85GB | 11.9 t/s
|
| 109 |
+
- Step 4660 | Loss: 9.5000 | VRAM: 14.85GB | 11.9 t/s
|
| 110 |
+
- Step 4665 | Loss: 10.8750 | VRAM: 14.84GB | 11.9 t/s
|
| 111 |
+
- Step 4670 | Loss: 11.0625 | VRAM: 14.84GB | 11.9 t/s
|
| 112 |
+
- Step 4675 | Loss: 10.8750 | VRAM: 14.84GB | 11.9 t/s
|
| 113 |
+
- Step 4680 | Loss: 9.2500 | VRAM: 14.84GB | 11.9 t/s
|
| 114 |
+
- Step 4685 | Loss: 9.0000 | VRAM: 14.85GB | 11.9 t/s
|
| 115 |
+
- Step 4690 | Loss: 10.5625 | VRAM: 14.84GB | 11.9 t/s
|
| 116 |
+
- Step 4695 | Loss: 10.1875 | VRAM: 14.84GB | 11.9 t/s
|
| 117 |
+
- Step 4700 | Loss: 8.6875 | VRAM: 14.85GB | 11.9 t/s
|
| 118 |
+
- Step 4705 | Loss: 10.7500 | VRAM: 14.85GB | 11.9 t/s
|
| 119 |
+
- Step 4710 | Loss: 9.2500 | VRAM: 14.84GB | 11.9 t/s
|
| 120 |
+
- Step 4715 | Loss: 8.3750 | VRAM: 14.84GB | 11.9 t/s
|
| 121 |
+
- Step 4720 | Loss: 9.9375 | VRAM: 14.84GB | 11.9 t/s
|
| 122 |
+
- Step 4725 | Loss: 10.8125 | VRAM: 14.84GB | 11.9 t/s
|
| 123 |
+
- Step 4730 | Loss: 9.8125 | VRAM: 14.84GB | 11.9 t/s
|
| 124 |
+
- Step 4735 | Loss: 9.2500 | VRAM: 14.84GB | 11.9 t/s
|
| 125 |
+
- Step 4740 | Loss: 10.6875 | VRAM: 14.84GB | 11.9 t/s
|
| 126 |
+
- Step 4745 | Loss: 10.1250 | VRAM: 14.84GB | 11.9 t/s
|
| 127 |
+
- Step 4750 | Loss: 10.6250 | VRAM: 14.84GB | 11.9 t/s
|
| 128 |
+
- Step 4755 | Loss: 10.8125 | VRAM: 14.85GB | 11.9 t/s
|
| 129 |
+
- Step 4760 | Loss: 10.4375 | VRAM: 14.85GB | 11.9 t/s
|
| 130 |
+
- Step 4765 | Loss: 10.3125 | VRAM: 14.84GB | 11.9 t/s
|
| 131 |
+
- Step 4770 | Loss: 9.5000 | VRAM: 14.84GB | 11.9 t/s
|
| 132 |
+
- Step 4775 | Loss: 10.0000 | VRAM: 14.85GB | 11.9 t/s
|
| 133 |
+
- Step 4780 | Loss: 9.5000 | VRAM: 14.85GB | 11.9 t/s
|
| 134 |
+
- Step 4785 | Loss: 9.0625 | VRAM: 14.84GB | 11.9 t/s
|