Home / Adreno (TM) 740 / Compare

Turnip-v26.3.0-20261013-r3 vs turnip_faux123_v1.1.9-rc14

On Adreno (TM) 740 (AYN Odin2, AYN Odin2 Portal, AYN Thor, HONOR Magic5 Pro, Moorechip Retroid Pocket 6, Moorechip Retroid Pocket Nova, motorola edge plus 2023, OnePlus 11 5G, POCO F6 Pro, Samsung Galaxy S23, Samsung Galaxy S23, Samsung Galaxy S23, Samsung Galaxy S23 Ultra, Samsung Galaxy S23 Ultra, Samsung Galaxy S23+, Sony Xperia 1 V, Xiaomi 13 pro). The first driver named is compared against the second, which is the baseline.

Stockturnip_faux123_v1.1.9-rc14Mesa 26.3.0-devel (git-dd45fd0baa)

CustomTurnip-v26.3.0-20261013-r3Mesa 26.3.0-devel (git-7b87f8ad22)

Graphics system version (Vulkan)1.4.359 to 1.4.363

Optional features (extensions)174 to 177+3 -0

Features that pass their check (validated)51 to 490 pass only on Turnip-v26.3.0-20261013-r3 2 pass only on turnip_faux123_v1.1.9-rc14

Speed test scenes (benchmark)0 faster 5 slower28% slower overall (geometric mean)

Games each driver could run (an estimate from its feature list): green likely, yellow maybe, orange limited, red no

StockDX92002+DX102006+DX112009+DX122015+Switch

CustomDX92002+DX102006+DX112009+DX122015+Switch

compute_nbody11% slower
hdr_2160p22% slower
oit_linked_lists33% slower
pipeline_barriers_2160p39% slower
subpasses_2160p33% slower

Custom driver against stock, per scene. Bars to the right are faster, to the left slower.

turnip_faux123_v1.1.9-rc14 details Turnip-v26.3.0-20261013-r3 details

Feature checks that differ (2)

Feature	turnip_faux123_v1.1.9-rc14	Turnip-v26.3.0-20261013-r3	Change
core.compute_udiv core	pass	fail	lost
core.copy_wide_row_length core	pass	fail	lost

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	turnip_faux123_v1.1.9-rc14 ms	Turnip-v26.3.0-20261013-r3 ms	Difference	Starts	GPU clock, MHz	Picture vs stock
compute_nbody	7.45 typical 7.49	8.40 typical 8.23	11% slower	3 of 3 vs 3 of 3	561 vs ? clock unknown	no large difference 0.0 no small difference worst 0.0 64x36
compute_nbody_2x	20.74 typical 20.77	n/a	only one side	3 of 3 vs ?	615 vs ? clock unknown	not checked
hdr_2160p	40.66 typical 39.92	52.19 typical 51.39	22% slower	3 of 3 vs 3 of 3	615 vs ? clock unknown	no large difference 0.0 no small difference worst 0.0 64x36
oit_linked_lists	26.27 typical 26.40	38.97 typical 38.81	33% slower	3 of 3 vs 3 of 3	636 vs ? clock unknown	no large difference 0.0 no small difference worst 0.0 64x36
pipeline_barriers	2.22 typical 2.18	3.72 typical 3.86	too light, see pipeline_barriers_2160p	3 of 3 vs 3 of 3	719 vs ? clock unknown	no large difference 0.0 no small difference worst 0.0 64x36
pipeline_barriers_2160p	8.56 typical 8.55	13.98 typical 13.84	39% slower	3 of 3 vs 3 of 3	615 vs ? clock unknown	no large difference 0.0 no small difference worst 0.0 64x36
subpasses	3.57 typical 3.57	5.94 typical 6.02	too light, see subpasses_2160p	3 of 3 vs 3 of 3	719 vs ? clock unknown	not checked
subpasses_2160p	14.86 typical 14.97	22.17 typical 22.12	33% slower	3 of 3 vs 3 of 3	615 vs ? 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 Turnip-v26.3.0-20261013-r3 against turnip_faux123_v1.1.9-rc14. "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)

Extensions only in Turnip-v26.3.0-20261013-r3 (3)

VK_KHR_shader_quad_control
VK_QCOM_rotated_copy_commands
VK_VALVE_buffer_device_address_allocation_alignment

Extensions only in turnip_faux123_v1.1.9-rc14 (0)

None

Core features that differ (0)

None

Limits that differ (0)

Limit	turnip_faux123_v1.1.9-rc14	Turnip-v26.3.0-20261013-r3

Formats supported by only one driver (0 gained, 0 lost)

Only in Turnip-v26.3.0-20261013-r3:

None

Only in turnip_faux123_v1.1.9-rc14:

None