A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 7, 16:30 UTC
[Bugfix] Fix uint32 overflow in Mamba selective scan state pointer arithmetic (#35275)
Signed-off-by: Josephasafg <ajgard7@gmail.com>
Asaf Gardin committed
Feb 26, 2026 at 14:22 UTC
ec13e549d3e1de13d05af759cc8bef3f7cf5e318
1 file changed
+1
-1
csrc/mamba/mamba_ssm/selective_scan.h
+1
-1
@@ -15,7 +15,7 @@
15
////////////////////////////////////////////////////////////////////////////////////////////////////
16
17
struct SSMParamsBase {
18
- using index_t = uint32_t;
18
+ using index_t = size_t;
19
20
int batch, dim, seqlen, dstate, n_groups, n_chunks;
21
int dim_ngroups_ratio;