Enable fp8 support (#1101)

* initial checkin

* resolve cr comments

* resolve the build issue

* fix the data correctless issue

* update fp8 header file and update the unit test for fp8 support

* remove fp16 from fp8 headers

* fix ut issue and catch up the latest code from develop

* udate according to cr comments

* update ut according to cr comments

* update num floats for each SumPostDiv from 4 to 6

* update fp8 header file name

* fix the typo

[ROCm/rccl commit: 6777e65c1d]
Tento commit je obsažen v:
Andy li
2024-03-09 07:17:53 +08:00
odevzdal GitHub
rodič 2354601589
revize e373bd44bf
29 změnil soubory, kde provedl 1243 přidání a 48 odebrání
+2 -2
Zobrazit soubor
@@ -32,7 +32,7 @@ namespace RcclUnitTesting
// Configuration
std::vector<ncclFunc_t> const funcTypes = {ncclCollReduce};
std::vector<ncclDataType_t> const dataTypes = {ncclFloat16, ncclFloat64};
std::vector<ncclDataType_t> const dataTypes = {ncclFloat16, ncclFloat64, ncclFp8E4M3, ncclFp8E5M2};
std::vector<ncclRedOp_t> const redOps = {ncclMin};
std::vector<int> const roots = {0};
std::vector<int> const numElements = {393216};
@@ -70,7 +70,7 @@ namespace RcclUnitTesting
// Configuration
std::vector<ncclFunc_t> const funcTypes = {ncclCollReduce};
std::vector<ncclDataType_t> const dataTypes = {ncclBfloat16};
std::vector<ncclDataType_t> const dataTypes = {ncclBfloat16, ncclFp8E4M3, ncclFp8E5M2};
std::vector<ncclRedOp_t> const redOps = {ncclMax};
std::vector<int> const roots = {0};
std::vector<int> const numElements = {393216};