[zz002]

Summary

SliceToHip now computes the true extent of a Slice whose start is -Shape(x)[k], instead of falling back to the data dim upper bound.

Related issue or design

None

Why

resolveHostIndex did not walk through Neg, so such a start failed to resolve and the extent fell back to the data dim. A graph that keeps the last S columns of the attention mask, Slice(attention_mask, -Shape(input_ids)[1], INT64_MAX), and gates a [B, S, ...] tensor with it via Where then gets a [B, total] mask at decode; Where broadcasts the [B, 1, ...] operand to [B, total, ...] and every downstream op sees the wrong sequence length. Prefill (S == total) hides it.

What

• resolveHostIndex resolves onnx.Neg and hip.neg as 0 - x. Both spellings are accepted for the same reason as Sub/Add/Mul: the greedy driver does not order this pattern against the producer rewrites.

Test plan

• LIT: test_slice.mlir tests 17 (ONNX spelling) and 18 (hip.neg spelling); full LIT suite passes
• Numeric: test_shape_ops.py::TestSlice::test_slice_mask_tail_by_sequence_length for (S, total) = (1, 20), (3, 12), (7, 7) matches CPU EP exactly; the two decode cases fail without the fix

Notes for reviewers

None

[github-actions]

Thanks for opening a PR!

This project follows LLVM's incremental-development and AI-tool-use
guidance. See CONTRIBUTING.md
for the project workflow.

Before requesting review, please check that:

1. The change is focused. Substantial work links the relevant issue
   or design discussion.
2. The PR documents relevant test results and updates affected
   documentation.
3. If AI tools provided substantial assistance, the description
   explains what was assisted and how it was validated, and commit
   trailers identify the tool. The contributor has reviewed and
   understands the result.

Reviewers are assigned through
CODEOWNERS where ownership
is configured.

[github-actions]

L2 Accuracy Results (EP vs CPU)

Model	Combined L2	Total Elems	Skipped NaN/Inf
conv_test_hybrid	1.12256E-07	64	0
GroupQueryAttention_seq256	59.2561	2621440	0
MatMulNBits_o_seq128	259.905	368640	0
QMoE_seq128	24.2583	368640	0

Threshold: 0.01 | Run: 7915 - Commit: db0f9a6

[github-actions]

Binary sizes

Baseline: run 40294400486 - 7579ebf (dll-sizes)

File	This (MB)	Main (MB)	Delta (MB)
hipgpu.dll	59.31	59.31	0.00
custom_kernels_gfx1170.dll	5.55	5.55	0.00
custom_kernels_gfx11-generic.dll	5.79	5.79	0.00

MorphiZen EP Performance Results

Model	QPS	Session (s)	1st Infer (ms)	CPU%	Mem (MB)
full_model_seq128	8.76	7.42	272	3	1158
GroupQueryAttention_seq128	11265.60	2.54773	1	5	229
matmul_down_seq128	524.12	3.30	119	3	274

EPContext Export Performance

Model	QPS	Session (s)	1st Infer (ms)	CPU%	Mem (MB)
full_model_seq128	8.82	47.95	246	3	15505

EPContext Import Performance

Model	QPS	Session (s)	1st Infer (ms)	CPU%	Mem (MB)
full_model_seq128	8.83	10.23	245	3	15638

OGA Benchmark Results

Model	Warmup	Reps	Prompt Len	Gen Tokens	TTFT (ms)	TPS	Peak Mem (GB)	GPU Mem (GB)
gpt-oss-20b-webgpu-int4-rtn-block-32	1	5	128	128	90.4	74.7	1.25	12.23
Llama-3.1-8B-awq-g128-int4-asym-fp16-onnx-dml	1	5	128	128	117.4	47.7	1.15	5.23

OGA Wheel Smoke (Python benchmark_e2e.py)

Model	TTFT (ms)	TPS
Llama-3.1-8B-awq-g128-int4-asym-fp16-onnx-dml	119	47.9

VLM Benchmark Results

Model	TTFT (ms)	TPS	Output
gemma-4-E4B-it-fp16-int4-kquant-prune-logits	3966.8	7.3	This is a high-quality, outdoor photograph featuring a beautiful, fluffy white dog, likely a West Highland White Terrier or a si
Qwen3.8-27B-fp16-ve-fp16-int4-k_quant-gs128-text-dml-onnx	14451.0	14.0	The user wants a detailed description of the image. Let me carefully observe all elements before writing. Subject: A white do

Run: 7915 - Commit: db0f9a6