A hands-on test training a MOSS-Image Base character LoRA at 1024px using ROCm PyTorch and ai-toolkit on a GMKtec EVO-X2 (Radeon 8060S). Covers adapting Anima captions to MOSS, tracking step times and VRAM usage up to 3,000 steps, and adjusting prompts and checkpoints.
Benchmarked Gemma 4 31B ARA-heretic on GMKtec EVO-X2 (Ryzen AI Max+ 395 / Radeon 8060S) using llama.cpp ROCm. While short prompts hit 8.4 t/s, 62K context runs collapsed into <unused49> and leaked across slots. Here is how upgrading the build and configuring --cache-ram resolved it.
The five official Anima setups took about 38 s (Turbo) and 260 s (Base) per image on an M1 Max 64GB. Under the same conditions on an EVO-X2 (Radeon 8060S, ROCm, Windows, ComfyUI), they took about 9 s and 65 s once loaded, and a character LoRA did not change those times.
A crash investigation log for the ASUS ROG Zephyrus G14 (Ryzen / RTX 3050 Ti Laptop). Analyzing MEMORY_MANAGEMENT errors in audio processing and GPU timeouts from crash dumps, setting up continuous logging, and removing the AMD driver.
Tested on EVO-X2 (Radeon 8060S): MiniMax H3 FL2VA first/last keyframe animation, 4.5s duo dialogue, 4-piece rock band, and reference audio pitch control across 3 seeds.
Claiming the $100 cloud credit from the AMD AI Developer Program involves a two-stage signup, a confusing DigitalOcean email loop, and non-obvious Credit balances. Here is the exact path through the UI.
Tested on EVO-X2 (gfx1151) under Windows + ROCm: the official b10666 binary runs it at pp512 159.25 / tg128 23.26, and the earlier crash was my own workaround flag.
Tested on GMKtec EVO-X2 (Ryzen AI Max+ 395): Q8_0 + MTP beats a 4-bit M1 Max at 22 tok/s, thinking burns 32,712 chars before any HTML, and the NSFW refusal line moves.
Tested local Wan video gen on a Radeon 8060S (Strix Halo, 48GB UMA, Windows). ZLUDA can't run stock PyTorch; AMD's TheRock gfx1151 wheel gives native ROCm. FastWan 1.3B in 4min, Wan 14B I2V in 13.6min — VAE decode and 16GB-RAM Segfaults are the real limits.
Based on EE Times' interview with AMD AI Software VP Anush Elangovan, we assess the ROCm vs CUDA ecosystem gap. Includes hands-on experience with ROCm breaking four times on Strix Halo, plus practical guidance on choosing between NVIDIA, AMD, and Apple Silicon.
Benchmarking NII's LLM-jp-4-32B-A3B-thinking on EVO-X2 (Ryzen AI Max+ 395) with ROCm. 62.9 t/s vs Qwen3.5-35B-A3B's 44.7 t/s. Covers thinking control issues, KV cache trade-offs, knowledge cutoff, Japanese quality comparisons, code generation tests, and training data composition.