[libc-commits] [libc] a1a610a - [libc] Increase the number of times we wait on a slab

[Joseph Huber via libc-commits]

Author: Joseph Huber
Date: 2025-07-28T09:23:29-05:00
New Revision: a1a610a1285fe4cde9f5f6a4a759da95266bdcb6

URL: https://github.com/llvm/llvm-project/commit/a1a610a1285fe4cde9f5f6a4a759da95266bdcb6
DIFF: https://github.com/llvm/llvm-project/commit/a1a610a1285fe4cde9f5f6a4a759da95266bdcb6.diff

LOG: [libc] Increase the number of times we wait on a slab

Summary:
This wait restricts how long we wait on a slab. The only reason this
isn't an infinite loop is to prevent complete deadlocks. However, this
limit was *just* on the cusp of waiting long enough for the allocation
to be done. Just increase this to a sufficiently large value, because
this limit only exists to keep the interface wait-free in the absolute
worst case scheduling scenario. This *MASSIVELY* improved performance
for mixed allocations as we no longer shuffled around creating more than
necessary.

Added: 
    

Modified: 
    libc/src/__support/GPU/allocator.cpp

Removed: 
    


################################################################################
diff  --git a/libc/src/__support/GPU/allocator.cpp b/libc/src/__support/GPU/allocator.cpp
index 8ec1003e6cb66..b9c86c5ab73c9 100644
--- a/libc/src/__support/GPU/allocator.cpp
+++ b/libc/src/__support/GPU/allocator.cpp
@@ -37,7 +37,7 @@ constexpr static uint32_t MIN_SIZE = 16;
 constexpr static uint32_t MIN_ALIGNMENT = MIN_SIZE - 1;
 
 // The number of times to attempt claiming an in-progress slab allocation.
-constexpr static uint32_t MAX_TRIES = 128;
+constexpr static uint32_t MAX_TRIES = 1024;
 
 // A sentinel used to indicate an invalid but non-null pointer value.
 constexpr static uint64_t SENTINEL = cpp::numeric_limits<uint64_t>::max();