nihui/ncnn - ncnn - 开源协同云脑生态支撑系统

Commit Graph

Author	SHA1	Message	Date
nihuini	1ea9de3bdf	create shader pipeline by type index, resolve binding count and push constant count from spirv. since we don't create compound shader module for macos and ios compatibility, it is enough to use fixed main as the shader entry point	6 years ago
nihui	f972bf49d1	enable bugihfa on rk3288 and rk3399	6 years ago
nihui	7c97142524	old qcom adreno driver seems to have the same bug as mali does	6 years ago
nihui	bbaa4dcce2	compile fp16pa, optimize shader for size, enable implicit fp16 arithmetic for qcom855 and qcom855plus	6 years ago
nihuini	b361b24832	do not enforce coherent memory type, queue transfer after uploading model weight	6 years ago
nihui	038666e049	the initial auto test (#1464 ) * cpu test * wip * ci run test * travis ci for arm64 * arm64 ctest * copy vulkan loader * wip * run * Update ccpp.yml * gpu test * swiftshader * cache macos swiftshader * try MoltenVK * try vulkaninfo * give swiftshader another try * disable failed macos gpu test * more conv test, fix conv3x3s1 gpu test fail * fix deconvolution test * dilation test * cmake option to build tests * ncnn_add_layer_test macro * host barrier before upload and after download, handle packing layout option * test packing layout * wip * wip * merge deconvolution packing and non-packing code * merge convolution packing and non-packing code * pass top_blob_count param * fix build * take care of non-coherent mappable memory	6 years ago
nihuini	a477aee0ba	print graphics queue info, const++	6 years ago
nihui	8a87f0267a	workaround local workgroup size specialization constant bug for old arm mali vulkan driver, fix #1424	6 years ago
nihui	a867d96822	dynamic memory type querying, respect memory requirement memory type bits	6 years ago
nihui	7e68c5e1e9	enable ycbcr conversion feature, get graphics queue	6 years ago
nihui	cb41b00e6e	setup VK_KHR_bind_memory2 functions	6 years ago
nihui	b29e8b0e09	check and enable more vulkan extensions	6 years ago
nihuini	21b5508c96	shared locked vkallocator cannot prevent concurrent accessing during actual gpu inference, use seperated vkallocator for each queue	7 years ago
nihuini	e9ffdb5bdd	16bit storage on arm mali is buggy	7 years ago
nihuini	040a8d2427	set vulkan device by gpu index	7 years ago
nihuini	5fdffbcaac	destroy_gpu_instance is not threadsafe anyway, fix deadlock on exit	7 years ago
nihuini	838c5df839	option api changes	7 years ago
nihuini	7f7bbf12e5	new api for getting the default gpu device	7 years ago
nihuini	cd7559c639	more fix for fp16p, still disabled by default	7 years ago
nihui	25b9736f82	shader fp16 packed	7 years ago
nihuini	4b50a97e31	implement vulkan winograd23	7 years ago
nihuini	37e150162a	do not retrieve timestamp availabitliy bits	7 years ago
nihuini	8e2fb2e710	expose timestamp_period and timestamp_valid_bits	7 years ago
nihuini	c9a9486307	merge command submit and wait, expose queue_count, concurrent queue submission shall work	7 years ago
nihuini	3d06c40d10	fix build with vulkan header version 65, fix #907	7 years ago
nihui	f92dcca3b3	compiled spirv nearly always claim uniform buffer 8bit / 16bit access capability	7 years ago
nihui	c180e87502	add compile shader module function, create pipeline from custom shader spv data	7 years ago
nihuini	31db9797df	interp bicubic shader, initialize mat member with zero	7 years ago
nihui	162c46647d	do not create fp16 shader module on unsupported platform	7 years ago
nihui	058bd65c88	fix fp16 shader creation	7 years ago
nihuini	4e3df863d5	fix enable feature pointer	7 years ago
nihuini	05bf09ba70	rename fp16_storage to support_fp16_storage	7 years ago
nihuini	332722af63	fix fp16a int8a exchange oops	7 years ago
nihuini	e59dc6fafe	proper usage of instance extension VK_KHR_get_physical_device_properties2, check fp16 and int8 feature	7 years ago
nihuini	d933f384b6	bump engine version	7 years ago
nihuini	038389fa63	blacklist known buggy driver	7 years ago
nihuini	5646b7d2c2	flatten image	7 years ago
nihuini	1f4bdd91b5	uint32_t typed workgroup size	7 years ago
nihuini	532054b453	expose more device info	7 years ago
nihuini	4ac56d3c1c	unified memory index is not mandatory, sanity check	7 years ago
nihuini	5f0ee22a33	treat as unified memory architecture if memory heap is same	7 years ago
nihui	10b8ac68cc	[WIP] vulkan compute (#618 ) * vulkan infrastructure * vkallocator and vkmat * layer interface for vulkan compute * wip... * default vulkan device, command wrapper, upload model weight in load_model to simplify layer interface * simplify command api, vkmat holds staging buffer, relu works * initialize specialization constant, simplify command dispatch, fix staging buffer copy with different shape, convolution works * init extension functions * dynamic local size and group count * group count=1 is invalid * regard device max workgroup size limit * fix relu oooops * decouple command record and staging allocation * create result blob * add pooling shader * buffer is faster than image :) * fix pooling shader * add innerproduct shader * readonly writeonly decoration * simplify buffer creation * decouple command and layer, VK_KHR_descriptor_update_template extension makes descriptor binding update easy :D * fix vulkan building issues in visual studio (#1) * fix building issues on visual studio * ignore benchmark * cancel changes * ... ... * decouple paramdict and vulkandevice * fix staging buffer destroy in model loading * remove vkdev member in option * add padding shader * simplify vulkan layer creation, simplify convolution and pooling shader for no padding, less debug output * add convolutiondepthwise and softmax shader * specialization float type, add leakyrelu * add dropout shader * add batchnorm shader * split vulkan forward * add scale shader * push constant type can be int or float * set_optimal_local_size_xyz * add eltwise shader * concat vulkan forward * fix convolution without bias * add dummy shader for concat and split, more fix ... * optional VK_KHR_descriptor_update_template and VK_KHR_push_descriptor * check VK_KHR_push_descriptor for vkCmdPushDescriptorSetWithTemplateKHR * binaryop and unaryop shader * hide raw command buffer * simple vkbenchncnn benchmark * create device with transfer queue * rename command to vkcompute, add vktransfer and layer upload_model interface * external VkMat, copy and map wrt buffer offset * command copy respect offset and size * decouple weight upload and load, simplify upload weight api, use one big staging buffer for uploading weights * fix build on android * binding count can not vary :( * barrier check state, fix sub-op destruction * declare local_size_xyz constant, fix crash on radv * fix local_size_xyz, second try * more barrier and state fix * fix softmax * reconstruct buffer memory allocator, reuse blob buffer, less verbose output * find unified memory type index * weight staging buffer allocator and weight buffer allocator, respect descriptor buffer offset alignment * use VK_KHR_descriptor_update_template for faster descriptor update if available, multithread pipeline creation * find more useful vulkan extensions and enable them * fix msvc build * respect VK_KHR_dedicated_allocation for weight buffer allocation * fix android build * fix bias name conflicts with metal * decouple pipeline and layer, building shader sources into shader module, dedicated create_pipeline api, simplify pipeline recording * drop dummy shader, inplace softmax, multiple shader module works * fix unique queue family index error * flatten support vulkan * mnasnet run * find shader module by name, each entry point per shader module, fix attribute/id conflict on moltenvk * some minor changes * add some high level api * use dedicated transfer queue to upload weight model * prefer mappable buffer on unified memory * global pooling and convolution fc, reuse staging buffer * implement ring-buffer style blob allocator, add VkBufferMemory capacity * use blob allocator for workspace blob, it works fine :) * vulkan option off * Update layer.cpp * fix build with vulkan off * less verbose output, fix crash on vulkan_compute off * merge benchncnn tool * allocator clear api, use new weight buffer allocator per net * add default locked allocator * mapped mat ptr api, persistent mapped memory works generally :) * travis ci linux vulkan * travis ci vulkan wip ... * more gpu wip ... * more gpu wip ... * wip... * wip... * wip... ... * wip... ios vulkan build... * find glslangValidator on ios build * use dynamic moltenvk library * travis ci wip ... * ios simulator does not support metal at all * fix cpu only extractor * optimize workgroup size, first try * optimize workgroup size, second try * conv1x1s1d1 vec4 * revert build system * fix ncnn2mem build * fix ncnn2mem build	7 years ago

42 Commits (1469bc8b19b83d44206f36abfa3dc7377feeef69)