Qualcomm_819.2_adpkg vs mainline-turnip-V30
On Adreno (TM) 650 (blackshark SHARK PAR-A0, MANGMI Pocket Max, Moorechip Retroid Pocket 5, POCO F4, Samsung Galaxy S20 FE, Samsung Galaxy S20 FE 5G, Sony Xperia 5 II). The first driver named is compared against the second, which is the baseline.
Stockmainline-turnip-V30Mesa 26.3.0-devel (git-9c51ede50f)
vsCustomQualcomm_819.2_adpkgDriver Build: 651341e91d, I8d46b3bbb7, 1754290135
Graphics system version (Vulkan)? to ?
Optional features (extensions)? to ?not checked on both
Features that pass their check (validated)not tested on bothmainline-turnip-V30 and Qualcomm_819.2_adpkg: no driver check yet
Speed test scenes (benchmark)3 faster 0 slower127% faster overall (geometric mean)
- oit_linked_lists129% faster
- pipeline_barriers109% faster
- subpasses145% faster
Custom driver against stock, per scene. Bars to the right are faster, to the left slower.
Small areas of the picture differ from the stock driver in 1 scene (pipeline_barriers: 1833 of 2304 cells). The picture matches overall, so this is a local rendering fault, not a different scene.
Did not start every time: compute_nbody (stock 0 of 3, custom 3 of 3), hdr_2160p (stock 0 of 3, custom 3 of 3).
mainline-turnip-V30 details Qualcomm_819.2_adpkg details
Feature checks that differ
mainline-turnip-V30 and Qualcomm_819.2_adpkg: no driver check (validation report) yet. Both drivers need a driver check before their feature checks can be compared.
Speed test scenes (benchmark)
The two drivers were not measured on the same set of device models. Frame times from different devices do not compare drivers, so read these numbers with that in mind.
| Scene | mainline-turnip-V30 ms | Qualcomm_819.2_adpkg ms | Difference | Starts | GPU clock, MHz | Picture vs stock |
|---|
| compute_nbody | n/a | 11.36 typical 11.33 | only one side | 0 of 3 vs 3 of 3 | ? vs 670 clock unknown | not checked |
| hdr_2160p | n/a | 79.51 typical 79.39 | only one side | 0 of 3 vs 3 of 3 | ? vs 670 clock unknown | not checked |
| oit_linked_lists | 76.42 typical 76.83 | 33.34 typical 33.35 | 129% faster | 3 of 3 vs 3 of 3 | ? vs 670 clock unknown | no large difference 0.0 no small difference worst 0.0 64x36 |
| pipeline_barriers | 10.22 typical 10.17 | 4.89 typical 4.74 | 109% faster | 3 of 3 vs 3 of 3 | ? vs 670 clock unknown | no large difference 10.2 differs in 1833 small areas worst 58.7 64x36 |
| pipeline_barriers_2160p | n/a | 17.11 typical 17.46 | only one side | ? vs 3 of 3 | ? vs 670 clock unknown | not checked |
| subpasses | 19.18 typical 19.13 | 7.83 typical 7.65 | 145% faster | 3 of 3 vs 3 of 3 | ? vs 670 clock unknown | not checked |
| subpasses_2160p | n/a | 29.07 typical 29.07 | only one side | ? vs 3 of 3 | ? vs 670 clock unknown | not checked |
Time to draw one frame, per scene: the average (mean) frame time of the scene's best round, which is how many frames the driver delivered; lower is faster. "Typical" beside it is the median frame, which can sit well under the average when a driver's frames come in two speeds. A pair of reports made before per-round averages were recorded is compared on the median. A scene whose rounds disagree by more than 15% is left out. A scene marked "paced" was held at that mean frame time by the driver or the display queue, together with at least one other scene of the same run; its time says how the frames were spaced, not how fast the driver rendered, so it is not compared. A scene marked "too light" took under 8 ms at the standard size (1920 by 1080) on both drivers, where the screen's frame queue spaces the frames more than the chip does; its heavier run is the one compared: the row ending in _2160p is the scene drawn at 3840 by 2160, and compute_nbody_2x is that scene with twice the particles. "Faster" and "slower" describe Qualcomm_819.2_adpkg against mainline-turnip-V30. "Picture vs stock" compares one fixed frame per scene as a coarse colour grid (0 is identical, above 40 is a different picture); it finds a scene that was not drawn. Where both reports carry the finer 64x36 grid, the second figure is how many of its 2,304 cells differ by more than 4 of 255, and "worst" is the largest single cell; this is what finds a small rendering error in a picture that otherwise matches. A row marked 16x9 comes from a report made before the finer grid existed. "Starts" is how many of a scene's process starts reached a measured round. A GPU clock marked "not comparable" differed by more than 5% of the device's maximum between the two timings.
Technical differences (Vulkan)
Both drivers need a driver check (validation report) before their technical details (Vulkan) can be compared.