{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":105299467,"defaultBranch":"master","name":"intel-graphics-compiler","ownerLogin":"intel","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2017-09-29T17:27:54.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/17888862?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1721133533.0","currentOid":""},"activityList":{"items":[{"before":"831268a56e0850247f8733ad33dd563a5c61a065","after":"cca46b8029177c10719198ea6b925b49e6160243","ref":"refs/heads/master","pushedAt":"2024-07-20T00:26:50.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" IGA: GED version update\n\nGED update to gedVersion = \"26602a3a\" for internal changes","shortMessageHtmlLink":" IGA: GED version update"}},{"before":"52cb3adb44f21d13810bd2a84d33ac608ca05a67","after":"831268a56e0850247f8733ad33dd563a5c61a065","ref":"refs/heads/master","pushedAt":"2024-07-19T22:08:30.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Extend pattern matching for inverse square root in VC\n\nVC should match `inv(sqrt(x))` and `sqrt(inv(x))` patterns and replace\nthem with `rsqrt(x)` to improve performance and accuracy.","shortMessageHtmlLink":" Extend pattern matching for inverse square root in VC"}},{"before":"3fc1e667be12e9c94b91f9086dc1925f7e9756ca","after":"52cb3adb44f21d13810bd2a84d33ac608ca05a67","ref":"refs/heads/master","pushedAt":"2024-07-19T20:09:15.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Revert of 'Introduction yaml files to define GenISA intrinsics'\n\nRevert of 'This change is to replace the GenISA intrinsics defined in a python script with these defined in an yaml file. This unlocks an opportunity to extend the way of describing their properties.'","shortMessageHtmlLink":" Revert of 'Introduction yaml files to define GenISA intrinsics'"}},{"before":"482b69743aa15334033edfa3c57e2e904e3e045c","after":"3fc1e667be12e9c94b91f9086dc1925f7e9756ca","ref":"refs/heads/master","pushedAt":"2024-07-19T19:09:28.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Do not apply GenXPatternMatch on scalar instructions\n\n.","shortMessageHtmlLink":" Do not apply GenXPatternMatch on scalar instructions"}},{"before":"4a09bac49a33cf10291d28b2d7789dbd7c77eb1a","after":"482b69743aa15334033edfa3c57e2e904e3e045c","ref":"refs/heads/master","pushedAt":"2024-07-19T18:08:45.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Add cm_cl builtin for fdiv_ieee\n\n.","shortMessageHtmlLink":" Add cm_cl builtin for fdiv_ieee"}},{"before":"bed3ead6d9087c9bf314924ee640f5eff6103065","after":"4a09bac49a33cf10291d28b2d7789dbd7c77eb1a","ref":"refs/heads/master","pushedAt":"2024-07-19T08:04:48.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Disable predicate promotion to scalar for fused EU in VC\n\nDue to the fused EU hardware restrictions, VC should promote predicate\noperations into vector operations instead of scalar ones.","shortMessageHtmlLink":" Disable predicate promotion to scalar for fused EU in VC"}},{"before":"04fd0e18e765a933d4d3eed06ec4aec5ee87e6cd","after":"bed3ead6d9087c9bf314924ee640f5eff6103065","ref":"refs/heads/master","pushedAt":"2024-07-19T01:32:26.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Fix the bug in the dst/src overlap checking in local RA\n\nWrong calcuation formula failed to check the potential overlap","shortMessageHtmlLink":" Fix the bug in the dst/src overlap checking in local RA"}},{"before":"c8c5eb16bf213e14fc7c69f51d2af128a797ae11","after":"04fd0e18e765a933d4d3eed06ec4aec5ee87e6cd","ref":"refs/heads/master","pushedAt":"2024-07-18T16:11:57.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Add pass checking if the address in load/store is valid.\n\nThis commit adds a new pass that iterate over all load/store\ninstructions and checks if the address used in the instruction\nis greater of equal than the minimum valid address defined\nby a user.","shortMessageHtmlLink":" Add pass checking if the address in load/store is valid."}},{"before":"04dc203a35d861de52deb333fb1d9d03189cc72e","after":"c8c5eb16bf213e14fc7c69f51d2af128a797ae11","ref":"refs/heads/master","pushedAt":"2024-07-18T10:07:40.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":"[Autobackout][FunctionalRegression]Revert of change: 65f49f80c71448d791865f1aec4c38e03a749c15: Enable internal LSC typed 2D intrinsics\n\n.","shortMessageHtmlLink":"[Autobackout][FunctionalRegression]Revert of change: <a class=\"commit-link\" data-hovercard-type=\"commit\" data-hovercard-url=\"https://github.com/intel/intel-graphics-compiler/commit/65f49f80c71448d791865f1aec4c38e03a749c15/hovercard\" href=\"https://github.com/intel/intel-graphics-compiler/commit/65f49f80c71448d791865f1aec4c38e03a749c15\"><tt>65f49f8</tt></a>: Enable …"}},{"before":"1283e0c62f6e3f128ae946861f4331e6952a1ed8","after":"04dc203a35d861de52deb333fb1d9d03189cc72e","ref":"refs/heads/master","pushedAt":"2024-07-18T04:32:52.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Fix DPAS peeling heuristics for VC\n\nThe current DPAS peeling heuristics intrinsic has threat non-DPAS\ninstructions as DPAS ones eligible for peeling. This patch fixes the\nheuristics to only consider DPAS instructions.","shortMessageHtmlLink":" Fix DPAS peeling heuristics for VC"}},{"before":"82a615689f320da00f7c03ad08dfcd1838dda1d4","after":"1283e0c62f6e3f128ae946861f4331e6952a1ed8","ref":"refs/heads/master","pushedAt":"2024-07-18T02:22:07.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" CS loacl id uniform\n\nOnly do uniform analysis for cs local IDs when it is linear walkOrder\nand ID layout is linear. Skip TileY and QuadTile as their semantice and\nimplmentation needs to be clarified first.","shortMessageHtmlLink":" CS loacl id uniform"}},{"before":"71165b18b9679adbed70c53149147a9f6ecfec92","after":"82a615689f320da00f7c03ad08dfcd1838dda1d4","ref":"refs/heads/master","pushedAt":"2024-07-18T01:34:37.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Pass to convert 2D block intrinsics into constituents.\n\nThe 2D block IO intrinsics are opaque to the compiler and identical\npayloads cannot be optimized by the compiler. Intrinsics have been\nintroduced to overcome this problem by allowing the 2D intrinsics to be\nbroken up into constituent stages (payload, set X and Y relative to\npayload, IO operation).\n\nThis PR adds a new pass, `Decompose2DBlockFuncs`, which converts the 2D\nIO intrinsics into the constituents, allowing subequent passes to\noptimize the code. At the current stage, the payload intrinsics can be\nhoisted above loops by LICM, and can be merged by EarlyCSA and others.\n\nThe `LSC2DBlockSetAddrPayloadField` intrinsics have a mode in which the\nIO location in memory is specified exactly (relative to the payload),\nand another mode in which the intrinsic increments the value. When a\npayload is used by the `LSC2DBlockSetAddrPayloadField` intrinsic in the\naccumulator mode, the payload cannot be hoisted. Because of this, I have\ngiven the `LSC2DBlockCreateAddrPayload` default attributes of\n`writeonly`, which is conservative for when the SetAddr intrinsic is in\naccumulator mode.\n\nThe `LSC2DBlockSetAddrPayloadField` intrinsic added by the\n`Decompose2DBlockFuncs` is always in the mode in which it explicitly\nspecifies the memory location. Since we have this guarantee, the related\npayload is described as not accessing memory, since the dependencies are\nall properly accounted for by the data dependency; this allows the\ndesired optimizations to function.\n\nThe 2D block intrinsics are not decomposed when they do not fall within\na loop, or if the payload would have loop-dependent parameters.","shortMessageHtmlLink":" Pass to convert 2D block intrinsics into constituents."}},{"before":"65f49f80c71448d791865f1aec4c38e03a749c15","after":"71165b18b9679adbed70c53149147a9f6ecfec92","ref":"refs/heads/master","pushedAt":"2024-07-18T00:26:39.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" IGA: Add swsb_encode_mode to commandline option\n\nAdd an option to set SWSB_ENCODE_MODE from IGA standalone tool's\ncommandline option.","shortMessageHtmlLink":" IGA: Add swsb_encode_mode to commandline option"}},{"before":"9dda44d975ac668d0c17b504248afc81a0b852bc","after":"65f49f80c71448d791865f1aec4c38e03a749c15","ref":"refs/heads/master","pushedAt":"2024-07-17T19:05:12.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Enable internal LSC typed 2D intrinsics\n\n.","shortMessageHtmlLink":" Enable internal LSC typed 2D intrinsics"}},{"before":"86502fabc58ff3273652a67237503163b4228eed","after":"9dda44d975ac668d0c17b504248afc81a0b852bc","ref":"refs/heads/master","pushedAt":"2024-07-17T18:07:33.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Add possibility to disable ShrinkArrayAlloca\n\nAdd a registry key and a compilation option to disable ShrinkArrayAlloca\npass.","shortMessageHtmlLink":" Add possibility to disable ShrinkArrayAlloca"}},{"before":"141008e7c54b5f1f2c048c6fe663d2a50731dd1d","after":"86502fabc58ff3273652a67237503163b4228eed","ref":"refs/heads/master","pushedAt":"2024-07-17T14:10:11.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" New GenISA intrinsic: WaveInterleave\n\nAdds new GenISA intrinsic WaveInterleave that does subgroup reduction on\neach n-th work item. For example, for SIMD8 and interleave step = 2,\nthe result is reduction of work items 0,2,4,6 and separate reduction of\nwork items 1,3,5,7.\n\nChange includes pattern match for interleave reduction implemented with\nsubgroup shuffles.","shortMessageHtmlLink":" New GenISA intrinsic: WaveInterleave"}},{"before":"1c20fe544ef6a81d90d7c91a8a8b749bf3369d4e","after":"141008e7c54b5f1f2c048c6fe663d2a50731dd1d","ref":"refs/heads/master","pushedAt":"2024-07-17T04:32:47.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Fix Cooperative Matrix prefetch resolving\n\njoint_matrix_prefetch spec was updated removing unused X and Y\ncoordinates parameters. Updated JointMatrixFuncsResolution\npass and lit tests to match new interface and make SYCL\ntest pass again.","shortMessageHtmlLink":" Fix Cooperative Matrix prefetch resolving"}},{"before":"46e4f0c7151d1c3c4565c1ea8cb2e25643e1aa74","after":"1c20fe544ef6a81d90d7c91a8a8b749bf3369d4e","ref":"refs/heads/master","pushedAt":"2024-07-17T03:07:30.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Purge MTP Opt. Disabling Code\n\nRecent modifications on this existing part are problematic, this change is to wipe out old code and prepare for a clean start.","shortMessageHtmlLink":" Purge MTP Opt. Disabling Code"}},{"before":"e707d372803d18e8d00882f7a809652654cfad2b","after":"46e4f0c7151d1c3c4565c1ea8cb2e25643e1aa74","ref":"refs/heads/master","pushedAt":"2024-07-17T02:19:08.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Support builtins for min and max operations in CMCL\n\n.","shortMessageHtmlLink":" Support builtins for min and max operations in CMCL"}},{"before":"761e55c8d13a140104303a934683e61718514802","after":"e707d372803d18e8d00882f7a809652654cfad2b","ref":"refs/heads/master","pushedAt":"2024-07-16T23:08:53.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Clean-up lit tests\n\n.","shortMessageHtmlLink":" Clean-up lit tests"}},{"before":"6a9cb7b9130b777e652312ecbbe94aec3e9c03c3","after":"761e55c8d13a140104303a934683e61718514802","ref":"refs/heads/master","pushedAt":"2024-07-16T18:08:17.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Add missing BMG enum to WA table\n\nAdd missing BMG enum to WA table","shortMessageHtmlLink":" Add missing BMG enum to WA table"}},{"before":"bdd9896ba54659cd125381fff0c8c9d48f514375","after":"6a9cb7b9130b777e652312ecbbe94aec3e9c03c3","ref":"refs/heads/master","pushedAt":"2024-07-16T14:07:57.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Introduction yaml files to define GenISA intrinsics\n\nThis change is to replace the GenISA intrinsics defined in a python script with these defined in an yaml file. This unlocks an opportunity to extend the way of describing their properties.","shortMessageHtmlLink":" Introduction yaml files to define GenISA intrinsics"}},{"before":"eed2993dc906c70476896b01b74b8a3a8a1bbea7","after":"fe135fbf5e1d898121ca9d1fdfb0086154beb6b4","ref":"refs/heads/releases/igc-1.0.17224","pushedAt":"2024-07-16T12:38:50.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"pszymich","name":"Pawel Szymichowski","path":"/pszymich","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/45690384?s=80&v=4"},"commit":{"message":" Revert barrier behavior.\n\nPrevious change caused small regressions in few testcases.\n\n(cherry picked from commit 1afff1c757e0d93a4f57575f395f37117ba83bf6)","shortMessageHtmlLink":" Revert barrier behavior."}},{"before":"55cb4a0add06b1d9f03aa648da3bbb55f0a33512","after":"eed2993dc906c70476896b01b74b8a3a8a1bbea7","ref":"refs/heads/releases/igc-1.0.17224","pushedAt":"2024-07-16T12:37:53.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"pszymich","name":"Pawel Szymichowski","path":"/pszymich","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/45690384?s=80&v=4"},"commit":{"message":" Fix SOA checkuser\n\nWhen handling bitcast from strTy to strTy, the users of bitcast must be\nchecked as well.\n\nThis change is to fix the missing check.\n\n(cherry picked from commit fa84ada73aa65523d3c7ae50366d161c7807b748)","shortMessageHtmlLink":" Fix SOA checkuser"}},{"before":"3ad9e4aa9073e2780eb28a17fa0ec243a9b88b9a","after":"bdd9896ba54659cd125381fff0c8c9d48f514375","ref":"refs/heads/master","pushedAt":"2024-07-16T11:09:05.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Extract ocloc binary name\n\nAt some point ocloc was renamed, also fix tools substitution","shortMessageHtmlLink":" Extract ocloc binary name"}},{"before":"3815f2207ecaa0f3244f5a46a6316e4cba689963","after":"55cb4a0add06b1d9f03aa648da3bbb55f0a33512","ref":"refs/heads/releases/igc-1.0.17224","pushedAt":"2024-07-16T09:10:46.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"pszymich","name":"Pawel Szymichowski","path":"/pszymich","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/45690384?s=80&v=4"},"commit":{"message":" Adjust BFN matching heuristics for pattern matching\n\nVC shouldn't match non-profitable patterns. Flag register cannot be\nsource for BFN instruction. The finalizer emits extra MOV instructions\nin such cases. So, it's better to avoid them, when the total amount of\ninstruction is not reduced.\n\nWe also need to reduce the use threshold for instructions folded into\nthe BFN. This will reduce the value live ranges and improve the\nregister pressure.\n\n(cherry picked from commit 24309c3d1c31fb62d600f023439186a8172fc6a1)","shortMessageHtmlLink":" Adjust BFN matching heuristics for pattern matching"}},{"before":"d79fbf457d5887635379a579a8fc0bebd2cfc5ed","after":"3815f2207ecaa0f3244f5a46a6316e4cba689963","ref":"refs/heads/releases/igc-1.0.17224","pushedAt":"2024-07-16T09:10:06.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"pszymich","name":"Pawel Szymichowski","path":"/pszymich","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/45690384?s=80&v=4"},"commit":{"message":"[Autobackout][FunctionalRegression]Revert of change: c41634112f8dc5a6b51511bf306e8869c9476ad9: Adjust BFN matching heuristics for pattern matching\n\nVC shouldn't match non-profitable patterns. Flag register cannot be\n    source for BFN instruction. The finalizer emits extra MOV instructions\n    in such cases. So, it's better to avoid them, when the total amount of\n    instruction is not reduced.\n\n    We also need to reduce the use threshold for instructions folded into\n    the BFN. This will reduce the value live ranges and improve the\n    register pressure.\n\n(cherry picked from commit 960d15ad581e3ad68aa6267646702f6c8d2e903e)","shortMessageHtmlLink":"[Autobackout][FunctionalRegression]Revert of change: <a class=\"commit-link\" data-hovercard-type=\"commit\" data-hovercard-url=\"https://github.com/intel/intel-graphics-compiler/commit/c41634112f8dc5a6b51511bf306e8869c9476ad9/hovercard\" href=\"https://github.com/intel/intel-graphics-compiler/commit/c41634112f8dc5a6b51511bf306e8869c9476ad9\"><tt>c416341</tt></a>: Adjust …"}},{"before":"f58188f18d25652f5f5d19e12cf3366e3e17026f","after":"3ad9e4aa9073e2780eb28a17fa0ec243a9b88b9a","ref":"refs/heads/master","pushedAt":"2024-07-16T04:34:09.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Translate genx block2d VC intrinsics into internal ones\n\n.","shortMessageHtmlLink":" Translate genx block2d VC intrinsics into internal ones"}},{"before":"eee9683dd279668ec6957a92412fc71e629735da","after":"f58188f18d25652f5f5d19e12cf3366e3e17026f","ref":"refs/heads/master","pushedAt":"2024-07-16T00:26:19.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Add LoopAllocaUpperbound pass\n\nThis pass pattern matches 1-BB loops with non-constant loop upper bound\nand memory access to alloca that is constant size array.\n\nThen if all alloca accesses go through inbounds GEPs we could create\na new loop count with constant upper bound of alloca array size.","shortMessageHtmlLink":" Add LoopAllocaUpperbound pass"}},{"before":"b920173724419667149081a6cba31331fbf4ce7a","after":"eee9683dd279668ec6957a92412fc71e629735da","ref":"refs/heads/master","pushedAt":"2024-07-15T22:06:45.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"VPG-SWE-Github","name":"VPG SWE Github","path":"/VPG-SWE-Github","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/34278352?s=80&v=4"},"commit":{"message":" Add Barrier Control Flow Optimization\n\nOptimize the barrier control flow to sync and flush on the running\nthread.","shortMessageHtmlLink":" Add Barrier Control Flow Optimization"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEhFyLrQA","startCursor":null,"endCursor":null}},"title":"Activity · intel/intel-graphics-compiler"}