5883 commits

Author	SHA1	Message	Date
Jason Volk	d377674748	ircd::simt: Split vector reduce_add to hadd.	2023-01-01 19:16:06 -08:00
Jason Volk	521a04730c	ircd:🆑 Conf items for device queueing and queue size defaults.	2023-01-01 21:28:54 +00:00
Jason Volk	517f109f3a	ircd:🆑:exec: Override local size from reqd_work_group_size attribute.	2022-11-07 01:51:19 +00:00
Jason Volk	231a5305ca	ircd:🆑 Additional device banner queries; improve formatting.	2022-11-03 18:28:53 +00:00
Jason Volk	110d4e7b17	ircd::gpt: Remove hostside backprop branch for now; simplify sample tokenizer.	2022-11-03 18:28:53 +00:00
Jason Volk	d5dc477de5	ircd: Minor fix makefile special recipe ordering.	2022-11-03 18:28:53 +00:00
Jason Volk	e8aea6f727	ircd:🆑 Optimize warp size query on AMD platforms supporting CL extension.	2022-11-03 18:28:53 +00:00
Jason Volk	35a994bafb	ircd:🆑 Tweak queue properties for profiling exclusive of device queues.	2022-11-03 18:28:53 +00:00
Jason Volk	db549a64ca	ircd::gpt: Add values for custom timestamp range to ctrl prof block.	2022-11-03 18:28:53 +00:00
Jason Volk	571271c6a6	ircd::gpt::pipe::code: Simplify source construction.	2022-11-03 18:28:53 +00:00
Jason Volk	601a7a3a6a	ircd:🆑:code: Add construction from files and directories.	2022-11-03 18:28:53 +00:00
Jason Volk	5aeadc2d52	ircd:🆑:code: Split createProgram related out of ctors; static iov_max.	2022-11-03 18:28:53 +00:00
Jason Volk	d95718ed6f	ircd::fs: Additional convenience tools for path suite.	2022-11-03 18:28:53 +00:00
Jason Volk	561be9973a	ircd::simt::norm: Barrier for overlapping input and output buffers. (gfx1011)	2022-11-03 18:28:53 +00:00
Jason Volk	071e108eac	ircd: Add amdgcn--amdhsa build recipes.	2022-11-03 18:28:53 +00:00
Jason Volk	ac163274f1	ircd::simt: Simplify keywording; tease out OpenCL version by target.	2022-10-30 18:44:50 +00:00
Jason Volk	e89703aa97	ircd::gpt::gpu: Limit dispatch to the number of control frame buffers.	2022-10-30 18:44:50 +00:00
Jason Volk	2609c21913	ircd::gpt::pipe: Enable mutable model; fixes for backpropagation; range stub.	2022-10-18 22:01:35 +00:00
Jason Volk	47117dde9a	ircd::gpt::model: Fixes for amdgpu mmaps; reenable mapping by default.	2022-10-16 21:12:29 +00:00
Jason Volk	1ac69df8cd	ircd::gpt::pipe::range: Use opts for construction.	2022-10-16 01:34:03 +00:00
Jason Volk	e8dad47247	ircd::gpt::model: Support mmap'ed and non-mmap'ed models based on conf.	2022-10-12 18:54:04 -07:00
Jason Volk	222d8404f1	ircd::server: Bail out of resolve handlers in case of unexpected callbacks.	2022-10-12 18:54:04 -07:00
Jason Volk	6a05fcefeb	ircd::simt: Consolidate timestamp counter sampling into inline.	2022-10-12 18:53:53 -07:00
Jason Volk	54e3b8f5b4	ircd::simt: Consolidate portables and macros from units into headers.	2022-10-11 02:13:46 +00:00
Jason Volk	831141727b	ircd::gpt::gpu: Add assert macro when trapping supported.	2022-10-09 03:29:29 +00:00
Jason Volk	c1168fcc30	ircd::gpt: Resolve behavior of opts.limit: 0=analysis, -n=unlimited, n=limited.	2022-10-09 03:13:09 +00:00
Jason Volk	6fb9cd7bf1	ircd::tokens: Relax noexcept on token result closures.	2022-10-08 15:37:14 -07:00
Jason Volk	a9b22c9052	ircd::lex: Handle js null as falsy boolean.	2022-10-08 10:49:41 -07:00
Jason Volk	4c44125d37	ircd::gpt::samp: Split dispatch branch into memfun.	2022-10-06 22:28:23 +00:00
Jason Volk	9682f406b3	ircd::gpt::gpu: Mute printf() on unsupporting platforms.	2022-10-06 22:26:19 +00:00
Jason Volk	2afaa4fa4f	ircd::gpt::pipe: Remove cruft.	2022-10-06 22:05:52 +00:00
Jason Volk	442dad869d	ircd::gpt: Resolve cycle count sampling; add debug log; fix count.	2022-10-06 22:01:41 +00:00
Jason Volk	0917a1f041	ircd::gpt::pipe: Resolve control page sync at ends of sample.	2022-10-06 18:54:29 +00:00
Jason Volk	be3fd4be62	ircd::gpt::pipe: Add 5th profile col for OpenCL 2.0; graceful default output.	2022-10-05 13:53:52 -07:00
Jason Volk	a9709568b4	ircd::gpt::pipe: Move the frame mapping command to the cycle.	2022-10-05 20:10:31 +00:00
Jason Volk	33afa8a4fc	ircd::gpt::gpu: Add global fence between attn and ffnn accumulations (gfx1011).	2022-10-05 20:10:31 +00:00
Jason Volk	c4cceb425c	ircd::gpt::gpu: Use explicit broadcast for local access.	2022-10-05 20:08:40 +00:00
Jason Volk	03e5f255c2	ircd:🆑 Modernize queue properties; enable ooe.	2022-10-05 20:08:40 +00:00
Jason Volk	331a417656	ircd::gpt::gpu: Fix keywording for OpenCL 2.0+.	2022-10-02 01:30:10 +00:00
Jason Volk	ae9f14bf86	ircd::gpt::pipe::code: Mute options for non-llvm pipe; attempt source build before bc.	2022-10-02 01:29:58 +00:00
Jason Volk	ef7471defd	ircd:🆑 Add debug logs for mem creations.	2022-10-01 22:06:45 +00:00
Jason Volk	736c8014dd	ircd:🆑 Default device info queries on init; promote log level; fix dev ref.	2022-09-30 17:30:55 +00:00
Jason Volk	3d990f0890	Revert "ircd:🆑 Isolate pthread enable hooking during init/finit." ... This reverts commit 1a29e591ef.	2022-09-30 17:30:44 +00:00
Jason Volk	35462afd04	ircd: Fix mods_ldso.o build flags as it includes boost/dll.	2022-09-30 17:30:39 +00:00
Jason Volk	da614e33a5	ircd:🆑 Replace niceness value with intensity by range partition.	2022-09-27 14:18:25 -07:00
Jason Volk	508d27eb40	ircd:🆑 Query to memoize warp sizes on init.	2022-09-27 12:23:43 -07:00
Jason Volk	e85ed0e0dd	ircd::gpt: Remove various cruft.	2022-09-24 16:40:39 -07:00
Jason Volk	4a8302038a	ircd::gpt::vocab: Add tokenization and detokenization count() convenience.	2022-09-21 16:42:07 -07:00
Jason Volk	2a608b8a7c	ircd:🆑 Clarify value in banner w/ pretty bytes.	2022-09-13 18:30:40 -07:00
Jason Volk	ae9f2c1199	ircd::fs: Add asio glue for io_uring read/write. (closes #37)	2022-09-10 12:00:54 -07:00