Skip to content

GitLab

Switch branch/tag
Find file Normal viewHistoryPermalink
example.cfg 2.67 KB
Newer Older
Gilbert Lee's avatar
Initial commit  
Gilbert Lee committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47 
# ConfigFile Format:
# ==================
# A Link is defined as a uni-directional transfer from src memory location to dst memory location
# executed by either CPU or GPU
# Each single line in the configuration file defines a set of Links to run in parallel

# There are two ways to specify the configuration file:

# 1) Basic
#    The basic specification assumes the same number of threadblocks/CUs used per GPU-executed Link
#    A positive number of Links is specified followed by that number of triplets describing each Link

#    #Links #CUs (srcMem1->Executor1->dstMem1) ... (srcMemL->ExecutorL->dstMemL)

# 2) Advanced
#    The advanced specification allows different number of threadblocks/CUs used per GPU-executed Link
#    A negative number of links is specified, followed by quadruples describing each Link
#    -#Links (srcMem1->Executor1->dstMem1 #CUs1) ... (srcMemL->ExecutorL->dstMemL #CUsL)

# Argument Details:
#   #Links  :   Number of Links to be run in parallel
#   #CUs    :   Number of threadblocks/CUs to use for a GPU-executed Link
#   srcMemL :   Source memory location (Where the data is to be read from). Ignored in memset mode
#   Executor:   Executor are specified by a character indicating executor type, followed by device index (0-indexed)
#               - C: CPU-executed  (Indexed from 0 to 1)
#               - G: GPU-executed  (Indexed from 0 to 3)
#   dstMemL :   Destination memory location (Where the data is to be written to)

#               Memory locations are specified by a character indicating memory type,
#               followed by device index (0-indexed)
#               Supported memory locations are:
#               - C:    Pinned host memory       (on NUMA node, indexed from 0 to [# NUMA nodes-1])
#               - B:    Fine-grain host memory   (on NUMA node, indexed from 0 to [# NUMA nodes-1])
#               - G:    Global device memory     (on GPU device indexed from 0 to [# GPUs - 1])
#               - F:    Fine-grain device memory (on GPU device indexed from 0 to [# GPUs - 1])

# Examples:
# 1 4 (G0->G0->G1)             Single link using 4 CUs on GPU0 to copy from GPU0 to GPU1
# 1 4 (C1->G2->G0)             Single link using 4 CUs on GPU2 to copy from CPU1 to GPU0
# 2 4 G0->G0->G1 G1->G1->G0    Runs 2 Links in parallel.  GPU0 to GPU1, and GPU1 to GPU0, each with 4 CUs
# -2 (G0 G0 G1 4) (G1 G1 G0 2) Runs 2 Links in parallel.  GPU0 to GPU1 with 4 CUs, and GPU1 to GPU0 with 2 CUs

# Round brackets and arrows' ->' may be included for human clarity, but will be ignored and are unnecessary
# Lines starting with # will be ignored. Lines starting with ## will be echoed to output

# Single GPU-executed link between GPUs 0 and 1 using 4 CUs
1 4 (G0->G0->G1)