FauxVkDB
English
Home / Adreno (TM) 825 / Compare

Qualcomm_849_adpkg vs Turnip_Gen8_V36

On Adreno (TM) 825 (Nothing Phone (3), OnePlus Nord 6, Oppo K13 Turbo Pro 5G, Poco F7, Poco F7, REDMI Turbo 4 Pro). The first driver named is compared against the second, which is the baseline.

StockTurnip_Gen8_V36Mesa 26.3.0-devel (git-c501e1d16e)
vs
CustomQualcomm_849_adpkgDriver Build: a2a1a29a71, Icb4aa5a425, 1758706229
Graphics system version (Vulkan)1.4.359 to 1.4.295
Optional features (extensions)176 to 174+25 -27
Features that pass their check (validated)49 to 441 pass only on Qualcomm_849_adpkg 6 pass only on Turnip_Gen8_V36
Speed test scenes (benchmark)3 faster 1 slower41% 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
  • hdr_2160p115% faster
  • oit_linked_lists65% faster
  • pipeline_barriers_2160p6% slower
  • subpasses_2160p19% 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 2 scenes (pipeline_barriers: 1712 of 2304 cells, pipeline_barriers_2160p: 1725 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).

Turnip_Gen8_V36 details Qualcomm_849_adpkg details

Feature checks that differ (24)

FeatureTurnip_Gen8_V36Qualcomm_849_adpkgChange
extendedDynamicState3ColorBlendEnable
VK_EXT_extended_dynamic_state3
passunsupportedlost
extendedDynamicState3ColorBlendEquation
VK_EXT_extended_dynamic_state3
passunsupportedlost
extendedDynamicState3ColorWriteMask
VK_EXT_extended_dynamic_state3
passunsupportedlost
extendedDynamicState3DepthClampEnable
VK_EXT_extended_dynamic_state3
passunsupportedlost
extendedDynamicState3LogicOpEnable
VK_EXT_extended_dynamic_state3
passunsupportedlost
extendedDynamicState3PolygonMode
VK_EXT_extended_dynamic_state3
passunsupportedlost
core.compute_udiv
core
failpassgained
extendedDynamicState3AlphaToCoverageEnable
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3AlphaToOneEnable
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3ConservativeRasterizationMode
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3DepthClipEnable
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3DepthClipNegativeOneToOne
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3ExtraPrimitiveOverestimationSize
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3LineRasterizationMode
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3ProvokingVertexMode
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3RasterizationSamples
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3RasterizationStream
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3SampleLocationsEnable
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3SampleMask
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
extendedDynamicState3TessellationDomainOrigin
VK_EXT_extended_dynamic_state3
untestedunsupportedchanged
shaderResourceMinLod
VkPhysicalDeviceFeatures
untestedunsupportedchanged
shaderResourceResidency
VkPhysicalDeviceFeatures
untestedunsupportedchanged
sparseResidency16Samples
VkPhysicalDeviceFeatures
unsupporteduntestedchanged
sparseResidencyAliased
VkPhysicalDeviceFeatures
untestedunsupportedchanged

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.

SceneTurnip_Gen8_V36 msQualcomm_849_adpkg msDifferenceStartsGPU clock, MHzPicture vs stock
compute_nbodyn/a10.13 typical 10.05only one side0 of 3 vs 3 of 3? vs ? clock unknownnot checked
hdr_2160p24.17 typical 23.3611.23 typical 11.04115% faster3 of 3 vs 3 of 3? vs ? clock unknownno large difference 0.0 no small difference worst 0.3 64x36
oit_linked_lists28.27 typical 28.6517.12 typical 16.7665% faster3 of 3 vs 3 of 3? vs ? clock unknownno large difference 0.0 no small difference worst 0.0 64x36
pipeline_barriersn/a2.46too light, see pipeline_barriers_2160p3 of 3 vs 3 of 3? vs ? clock unknownno large difference 10.8 differs in 1712 small areas worst 51.7 64x36
pipeline_barriers_2160p8.59 typical 8.389.15 typical 9.176% slower3 of 3 vs 3 of 3? vs ? clock unknownno large difference 12.0 differs in 1725 small areas worst 65.3 64x36
subpasses4.70 typical 4.653.97 typical 3.55too light, see subpasses_2160p3 of 3 vs 3 of 3? vs ? clock unknownnot checked
subpasses_2160p17.66 typical 17.4814.89 typical 14.8319% faster3 of 3 vs 3 of 3? vs ? clock unknownnot 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_849_adpkg against Turnip_Gen8_V36. "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 Qualcomm_849_adpkg (25)
  • VK_ANDROID_external_format_resolve
  • VK_EXT_astc_decode_mode
  • VK_EXT_blend_operation_advanced
  • VK_EXT_depth_range_unrestricted
  • VK_EXT_device_fault
  • VK_EXT_external_memory_acquire_unmodified
  • VK_EXT_fragment_density_map2
  • VK_EXT_image_compression_control_swapchain
  • VK_EXT_pipeline_protected_access
  • VK_EXT_pipeline_robustness
  • VK_EXT_shader_subgroup_ballot
  • VK_EXT_subpass_merge_feedback
  • VK_KHR_pipeline_binary
  • VK_KHR_ray_tracing_position_fetch
  • VK_KHR_shader_maximal_reconvergence
  • VK_KHR_shader_quad_control
  • VK_KHR_shared_presentable_image
  • VK_NV_optical_flow
  • VK_QCOM_filter_cubic_weights
  • VK_QCOM_image_processing2
  • VK_QCOM_render_pass_store_ops
  • VK_QCOM_render_pass_transform
  • VK_QCOM_rotated_copy_commands
  • VK_QCOM_tile_properties
  • VK_QCOM_ycbcr_degamma
Extensions only in Turnip_Gen8_V36 (27)
  • VK_AMD_buffer_marker
  • VK_ARM_rasterization_order_attachment_access
  • VK_EXT_attachment_feedback_loop_dynamic_state
  • VK_EXT_attachment_feedback_loop_layout
  • VK_EXT_custom_resolve
  • VK_EXT_debug_marker
  • VK_EXT_depth_clamp_control
  • VK_EXT_dynamic_rendering_unused_attachments
  • VK_EXT_extended_dynamic_state3
  • VK_EXT_external_memory_dma_buf
  • VK_EXT_fragment_density_map_offset
  • VK_EXT_image_drm_format_modifier
  • VK_EXT_legacy_dithering
  • VK_EXT_legacy_vertex_attributes
  • VK_EXT_memory_budget
  • VK_EXT_nested_command_buffer
  • VK_EXT_primitive_restart_index
  • VK_EXT_shader_replicated_composites
  • VK_EXT_shader_uniform_buffer_unsized_array
  • VK_EXT_zero_initialize_device_memory
  • VK_GOOGLE_decorate_string
  • VK_GOOGLE_hlsl_functionality1
  • VK_KHR_maintenance7
  • VK_KHR_shader_relaxed_extended_instruction
  • VK_NV_compute_shader_derivatives
  • VK_VALVE_fragment_density_map_layered
  • VK_VALVE_mutable_descriptor_type
Core features that differ (5)
multiviewGeometryShader Vulkan 1.1pass to unsupported
protectedMemory Vulkan 1.1unsupported to pass
shaderInputAttachmentArrayDynamicIndexing Vulkan 1.2unsupported to pass
shaderInputAttachmentArrayNonUniformIndexing Vulkan 1.2unsupported to pass
pipelineProtectedAccess Vulkan 1.4unsupported to pass
Limits that differ (22)
LimitTurnip_Gen8_V36Qualcomm_849_adpkg
maxTexelBufferElements536870911134217728
bufferImageGranularity641
sparseAddressSpaceSize2748779069444294967295
maxPerStageResources1677721650331648
maxDescriptorSetUniformBuffersDynamic1632
maxDescriptorSetStorageBuffersDynamic816
maxVertexInputAttributeOffset40954096
maxGeometryInputComponents64128
maxGeometryOutputVertices2561024
maxGeometryTotalOutputComponents1024131072
maxFragmentInputComponents128112
maxComputeSharedMemorySize6553632768
maxComputeWorkGroupInvocations20481024
minMemoryMapAlignment409664
minUniformBufferOffsetAlignment64256
minStorageBufferOffsetAlignment464
maxFramebufferLayers10242048
discreteQueuePriorities24
pointSizeRange1, 40921, 4095
optimalBufferCopyOffsetAlignment12864
optimalBufferCopyRowPitchAlignment12864
nonCoherentAtomSize641
Formats supported by only one driver (3 gained, 5 lost)

Only in Qualcomm_849_adpkg:

  • G10X6_B10X6R10X6_2PLANE_420_UNORM_3PACK16
  • G10X6_B10X6R10X6_2PLANE_422_UNORM_3PACK16
  • G16_B16R16_2PLANE_420_UNORM

Only in Turnip_Gen8_V36:

  • B8G8R8A8_SSCALED
  • B8G8R8A8_USCALED
  • B8G8R8_SINT
  • B8G8R8_SNORM
  • B8G8R8_UINT