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
|
From: Vasimuddin Md <wasim.mzr@gmail.com>
Date: Thu, 9 May 2024 00:19:51 +0530
Subject: Merge pull request #229 from gh-jphan/fix_n_processed
Origin: upstream, https://github.com/bwa-mem2/bwa-mem2/commit/7aa5ff6c3330490e5629ab9b7327683d2dce02d6
Forwarded: not-needed
Fix threading reproducibility issue by moving n_processed update line from step 2 to step 1
---
src/fastmap.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/fastmap.cpp b/src/fastmap.cpp
index baf4a27..5cbb2dc 100644
--- a/src/fastmap.cpp
+++ b/src/fastmap.cpp
@@ -295,13 +295,13 @@ ktp_data_t *kt_pipeline(void *shared, int step, void *data, mem_opt_t *opt, work
w);
}
tprof[MEM_PROCESS2][0] += __rdtsc() - tim;
-
+
+ aux->n_processed += ret->n_seqs;
return ret;
}
/* Step 3: Write output */
else if (step == 2)
{
- aux->n_processed += ret->n_seqs;
uint64_t tim = __rdtsc();
for (int i = 0; i < ret->n_seqs; ++i)
|
