347 Commits

Author	SHA1	Message	Date
Peter Boyle	81eb1fd9f2	Accelerator loop changes for coalesced access	2019-06-15 07:44:47 +01:00
Peter Boyle	cb93d32cd9	Thread loop changes	2019-06-15 07:44:08 +01:00
Peter Boyle	8f223962ff	Thread loop changed	2019-06-15 07:43:42 +01:00
Peter Boyle	36f06555a2	Simplify Impl	2019-06-09 22:26:27 +01:00
Peter Boyle	d6c0e0756d	Remove GPU version	2019-06-09 11:23:42 +01:00
Peter Boyle	3e41b1055c	Remove Gpu only kernels.	2019-06-09 11:20:01 +01:00
Peter Boyle	e78a5e7838	ASM instantiation without link errors	2019-06-09 01:25:21 +01:00
Peter Boyle	8e3a05d89b	Moving the instantiation into a cleaner structure	2019-06-08 13:48:33 +01:00
Peter Boyle	8adc5da7dd	Testig out approaches to kernel writing introducing SIMT_loop temporarily	2019-06-08 13:47:04 +01:00
Peter Boyle	29a244e423	Test of using a lane variable instead of repeated reference to threadIdx.y	2019-06-08 13:46:26 +01:00
Peter Boyle	c933ac2248	Temporarily introduce a SIMT_loop to test out approaches prior to making a global change to ... accelerator_loop	2019-06-08 13:44:27 +01:00
Peter Boyle	ad2c433574	Instantiations move. Tried using Gianluca's suggestion about avoiding threadIdx but doesn't ... seem to make a difference. Will revisit this and probably remove the lane parameter from the coalescedRead	2019-06-08 13:43:12 +01:00
Peter Boyle	86e7fb6e86	Instantiation relocation	2019-06-08 13:42:46 +01:00
Peter Boyle	fb91dda7be	Hand instantiation moved location	2019-06-08 13:42:26 +01:00
Peter Boyle	82cf7bc5ab	Move instantiation into fermion/instantiation	2019-06-08 13:41:46 +01:00
Peter Boyle	e452cc0a22	Move static variables into instantiation .cc file	2019-06-08 13:41:20 +01:00
Peter Boyle	4d2b938166	Remove explict instantiation from here	2019-06-08 13:41:01 +01:00
Peter Boyle	10d16ab76c	Remove explict instantiation from here	2019-06-08 13:40:32 +01:00
Peter Boyle	1f997fa484	Instantiate via explict .cc files for parallel make.	2019-06-08 13:39:51 +01:00
Peter Boyle	dc5024e88c	The GPU reduction was not working for me and causing errors. Need to revisit. ... Gianluca is working on deterministic reduction/	2019-06-08 13:39:11 +01:00
Peter Boyle	0ee6e77cbc	Compiles GPU and CPU, still gives good performance on CPU	2019-06-05 13:28:16 +01:00
Peter Boyle	18d3cde29a	Compile on GPU workd	2019-06-05 00:14:58 +01:00
Peter Boyle	7323099966	Instatiation fix	2019-06-05 00:14:38 +01:00
Peter Boyle	6379651cdd	Generic or GPU ready for benchmark test on GPU	2019-06-05 00:13:52 +01:00
Peter Boyle	ba4fd756b9	Fix signature, but deprecating this loops style	2019-06-05 00:12:36 +01:00
Peter Boyle	d185fc1ebf	clean up instantiation	2019-06-05 00:11:52 +01:00
Peter Boyle	96b36d8367	Instantiation clean up	2019-06-05 00:11:27 +01:00
Peter Boyle	899f8b5065	Instantiation clean up 5d vec removal	2019-06-05 00:11:05 +01:00
Peter Boyle	c8d0483fe9	Remove 5d vectorisation	2019-06-05 00:10:37 +01:00
Peter Boyle	0f214e5f76	Clean up instantiation	2019-06-05 00:10:13 +01:00
Peter Boyle	8eea568426	GPU loop ; presently differentiated with ifdef, find a way to unify.	2019-06-05 00:09:28 +01:00
Peter Boyle	9636324069	GPU happy code	2019-06-05 00:08:54 +01:00
Peter Boyle	8a5489d9e6	Move the loop into a central kernel call.	2019-06-05 00:08:13 +01:00
Peter Boyle	8113845f9c	coalesce loop. Need to rationalise this file	2019-06-04 23:49:29 +01:00
Peter Boyle	b47f73c222	GPU happy	2019-06-04 21:30:39 +01:00
Peter Boyle	5720ced0fd	Simplifying	2019-06-04 21:30:08 +01:00
Peter Boyle	2c87b56b53	Making GPU happier	2019-06-04 21:29:44 +01:00
Peter Boyle	dbad48d802	Remove Ls vectorised DWF	2019-06-04 21:27:40 +01:00
Peter Boyle	4557a1365a	Remove Ls vectorised DWF	2019-06-04 20:59:59 +01:00
Peter Boyle	16e9b87d98	Remove Ls vectorised DWF as unused and hard to maintain	2019-06-04 20:59:01 +01:00
Peter Boyle	685eea3d0f	Small cosmetic	2019-06-04 20:58:14 +01:00
Peter Boyle	65b48831fb	Simplify code	2019-06-04 20:56:30 +01:00
Peter Boyle	57396fc595	Simplify code	2019-06-04 20:56:23 +01:00
Peter Boyle	a2e199df50	Simplifying Cayley cases.	2019-06-04 20:54:52 +01:00
Peter Boyle	c2625a127e	Non blocking loop. Want to change the naming here.	2019-06-04 20:52:59 +01:00
Peter Boyle	8794d35c78	GPU	2019-06-04 20:52:27 +01:00
Peter Boyle	24bff6dbe6	Minor improvements	2019-06-04 20:51:48 +01:00
Peter Boyle	45b15d10d3	GPU happy changes	2019-06-04 20:49:16 +01:00
Peter Boyle	33d6bbe32b	GPU must use accelerator vectors	2019-06-04 20:48:52 +01:00
Peter Boyle	7a1569bd46	Annoying, cannot rely on equivalence of Grid ComplexD adn Eigen Complex type on GPU. ... Solve with ComplexD typecasts but must be a better way	2019-06-04 20:47:49 +01:00