fix: skip buffer validation if on device

[0ax1]

No description provided.

[codspeed-hq]

Merging this PR will degrade performance by 30.15%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚡ 5 improved benchmarks
❌ 16 regressed benchmarks
✅ 1158 untouched benchmarks
⏩ 1323 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	bench_many_codes_few_values[2048]	369.1 µs	332.9 µs	+10.89%
⚡	Simulation	bench_many_codes_few_values[4096]	375.2 µs	338.9 µs	+10.7%
⚡	Simulation	bench_sparse_coverage[0.1]	366.8 µs	330.5 µs	+10.97%
⚡	Simulation	bench_sparse_coverage[0.5]	365.5 µs	329.2 µs	+11.01%
⚡	Simulation	bench_many_codes_few_values[1024]	366.5 µs	330.3 µs	+10.98%
❌	Simulation	canonical_into_non_nullable[(10000, 10, 0.0)]	205.4 µs	279.2 µs	-26.44%
❌	Simulation	canonical_into_non_nullable[(10000, 10, 0.01)]	232.6 µs	306.7 µs	-24.15%
❌	Simulation	canonical_into_non_nullable[(10000, 10, 0.1)]	391.9 µs	465.7 µs	-15.85%
❌	Simulation	into_canonical_non_nullable[(10000, 10, 0.01)]	234.9 µs	307.2 µs	-23.56%
❌	Simulation	canonical_into_non_nullable[(10000, 100, 0.01)]	2.1 ms	3 ms	-27.76%
❌	Simulation	canonical_into_non_nullable[(10000, 100, 0.0)]	1.9 ms	2.7 ms	-29.98%
❌	Simulation	into_canonical_nullable[(10000, 10, 0.1)]	635.2 µs	714.9 µs	-11.15%
❌	Simulation	into_canonical_non_nullable[(10000, 10, 0.0)]	207.5 µs	280.1 µs	-25.94%
❌	Simulation	into_canonical_nullable[(10000, 10, 0.0)]	461.8 µs	537.8 µs	-14.14%
❌	Simulation	into_canonical_non_nullable[(10000, 100, 0.01)]	2.2 ms	3 ms	-27.57%
❌	Simulation	into_canonical_non_nullable[(10000, 100, 0.1)]	3.7 ms	4.5 ms	-17.44%
❌	Simulation	canonical_into_non_nullable[(10000, 100, 0.1)]	3.7 ms	4.5 ms	-17.45%
❌	Simulation	into_canonical_nullable[(10000, 100, 0.1)]	6.1 ms	6.9 ms	-11.51%
❌	Simulation	into_canonical_nullable[(10000, 100, 0.0)]	4.4 ms	5.1 ms	-13.29%
❌	Simulation	into_canonical_non_nullable[(10000, 10, 0.1)]	393.9 µs	469.8 µs	-16.15%
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

_{Comparing ad/skip-device-buf-validation (60bd955) with develop (cdf10a6)}

Footnotes

1. 1323 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[joseph-isaacs]

this is wrong

[joseph-isaacs]

we should validate anything on the CPU and maybe in debug move to CPU and check?

[joseph-isaacs]

move to CPU and validate on CPU in debug mode?