Home / Mali-G77 / Compare

EXP_PanVK-kbase-rp4pro-lab86 vs EXP_panvk_rp4pro_lab17f

On Mali-G77 (Infinix GT 10 Pro, motorola edge 40, Redmi K40 Gaming, Redmi Note 10 Pro, Retroid Pocket 4 Pro, Samsung Galaxy Note20 Ultra, Samsung Galaxy S20 Ultra 5G). The first driver named is compared against the second, which is the baseline.

StockEXP_panvk_rp4pro_lab17fMesa 26.3.0-devel (git-eadd5fabd8)

CustomEXP_PanVK-kbase-rp4pro-lab86PanVK-kbase RC3-lowmem (Mesa 26.3.0-devel (git-a6db27309d))

Graphics system version (Vulkan)1.3.354 to 1.3.363

Optional features (extensions)166 to 175+12 -3

Features that pass their check (validated)49 to 523 pass only on EXP_PanVK-kbase-rp4pro-lab86 0 pass only on EXP_panvk_rp4pro_lab17f

Speed test scenes (benchmark)4 faster 0 slower 1 even14% faster 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_nbody5% faster
hdr_2160p17% faster
oit_linked_lists1% faster
pipeline_barriers50% faster
subpasses3% 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: 1574 of 2304 cells). The picture matches overall, so this is a local rendering fault, not a different scene.

EXP_panvk_rp4pro_lab17f details EXP_PanVK-kbase-rp4pro-lab86 details

Feature checks that differ (6)

Feature	EXP_panvk_rp4pro_lab17f	EXP_PanVK-kbase-rp4pro-lab86	Change
robustImageAccess2 VK_EXT_robustness2	unsupported	pass	gained
compat.helper_loop_terminator core	untested	pass	gained
compat.nested_map core	untested	pass	gained
extendedDynamicState2PatchControlPoints VK_EXT_extended_dynamic_state2	unsupported	untested	changed
depthBounds VkPhysicalDeviceFeatures	unsupported	untested	changed
shaderTessellationAndGeometryPointSize VkPhysicalDeviceFeatures	untested	unsupported	changed

Speed test scenes (benchmark)

Scene	EXP_panvk_rp4pro_lab17f ms	EXP_PanVK-kbase-rp4pro-lab86 ms	Difference	Starts	GPU clock, MHz	Picture vs stock
compute_nbody	16.47 typical 16.04	15.69 typical 15.60	5% faster	3 of 3 vs 3 of 3	827 vs 831	no large difference 0.0 no small difference worst 0.0 64x36
hdr_2160p	42.14 typical 42.04	35.88 typical 35.86	17% faster	3 of 3 vs 3 of 3	836 vs 828	no large difference 0.0 no small difference worst 0.7 64x36
oit_linked_lists	23.72 typical 23.69	23.46 typical 23.32	1% faster	3 of 3 vs 3 of 3	836 vs 836	no large difference 0.0 no small difference worst 0.0 64x36
pipeline_barriers	10.67 typical 10.80	7.10 typical 7.02	50% faster	3 of 3 vs 3 of 3	836 vs 836	no large difference 9.0 differs in 1574 small areas worst 66.0 64x36
pipeline_barriers_2160p	n/a	25.50 typical 25.59	only one side	? vs 3 of 3	? vs 836 clock unknown	not checked
subpasses	19.45 typical 19.47	18.82 typical 18.81	3% faster	3 of 3 vs 3 of 3	824 vs 834	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 EXP_PanVK-kbase-rp4pro-lab86 against EXP_panvk_rp4pro_lab17f. "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 EXP_PanVK-kbase-rp4pro-lab86 (12)

VK_EXT_conditional_rendering
VK_EXT_image_compression_control
VK_EXT_present_timing
VK_EXT_primitives_generated_query
VK_EXT_swapchain_maintenance1
VK_KHR_present_id
VK_KHR_present_id2
VK_KHR_present_wait
VK_KHR_present_wait2
VK_KHR_shader_constant_data
VK_KHR_swapchain_maintenance1
VK_KHR_swapchain_mutable_format

Extensions only in EXP_panvk_rp4pro_lab17f (3)

VK_KHR_depth_clamp_zero_one
VK_KHR_pipeline_binary
VK_KHR_robustness2

Core features that differ (1)

shaderOutputViewportIndex Vulkan 1.2

unsupported to pass

Limits that differ (7)

Limit	EXP_panvk_rp4pro_lab17f	EXP_PanVK-kbase-rp4pro-lab86
sparseAddressSpaceSize	0	140737454800896
maxGeometryShaderInvocations	32	64
maxGeometryInputComponents	128	64
maxGeometryOutputVertices	1024	256
minStorageBufferOffsetAlignment	16	4
timestampComputeAndGraphics	0	1
timestampPeriod	0	76.9231

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

Only in EXP_PanVK-kbase-rp4pro-lab86:

None

Only in EXP_panvk_rp4pro_lab17f:

None