From: Richard Henderson <rth@twiddle.net>
To: qemu-devel@nongnu.org
Cc: peter.maydell@linaro.org, "Emilio G. Cota" <cota@braap.org>
Subject: [Qemu-devel] [PULL 25/35] target-i386: remove helper_lock()
Date: Sat, 22 Oct 2016 14:05:14 -0700 [thread overview]
Message-ID: <1477170324-17783-26-git-send-email-rth@twiddle.net> (raw)
In-Reply-To: <1477170324-17783-1-git-send-email-rth@twiddle.net>
From: "Emilio G. Cota" <cota@braap.org>
It's been superseded by the atomic helpers.
The use of the atomic helpers provides a significant performance and scalability
improvement. Below is the result of running the atomic_add-test microbenchmark with:
$ x86_64-linux-user/qemu-x86_64 tests/atomic_add-bench -o 5000000 -r $r -n $n
, where $n is the number of threads and $r is the allowed range for the additions.
The scenarios measured are:
- atomic: implements x86' ADDL with the atomic_add helper (i.e. this patchset)
- cmpxchg: implement x86' ADDL with a TCG loop using the cmpxchg helper
- master: before this patchset
Results sorted in ascending range, i.e. descending degree of contention.
Y axis is Throughput in Mops/s. Tests are run on an AMD machine with 64
Opteron 6376 cores.
atomic_add-bench: 5000000 ops/thread, [0,1] range
25 ++---------+----------+---------+----------+----------+----------+---++
+ atomic +-E--+ + + + + + |
|cmpxchg +-H--+ |
20 +Emaster +-N--+ ++
|| |
|++ |
|| |
15 +++ ++
|N| |
|+| |
10 ++| ++
|+|+ |
| | -+E+------ +++ ---+E+------+E+------+E+-----+E+------+E|
|+E+E+- +++ +E+------+E+-- |
5 ++|+ ++
|+N+H+--- +++ |
++++N+--+H++----+++ + +++ --++H+------+H+------+H++----+H+---+--- |
0 ++---------+-----H----+---H-----+----------+----------+----------+---H+
0 10 20 30 40 50 60
Number of threads
atomic_add-bench: 5000000 ops/thread, [0,2] range
25 ++---------+----------+---------+----------+----------+----------+---++
++atomic +-E--+ + + + + + |
|cmpxchg +-H--+ |
20 ++master +-N--+ ++
|E| |
|++ |
||E |
15 ++| ++
|N|| |
|+|| ---+E+------+E+-----+E+------+E|
10 ++| | ---+E+------+E+-----+E+--- +++ +++
||H+E+--+E+-- |
|+++++ |
| || |
5 ++|+H+-- +++ ++
|+N+ - ---+H+------+H+------ |
+ +N+--+H++----+H+---+--+H+----++H+--- + + +H+---+--+H|
0 ++---------+----------+---------+----------+----------+----------+---++
0 10 20 30 40 50 60
Number of threads
atomic_add-bench: 5000000 ops/thread, [0,8] range
40 ++---------+----------+---------+----------+----------+----------+---++
++atomic +-E--+ + + + + + |
35 +cmpxchg +-H--+ ++
| master +-N--+ ---+E+------+E+------+E+-----+E+------+E|
30 ++| ---+E+-- +++ ++
| | -+E+--- |
25 ++E ---- +++ ++
|+++++ -+E+ |
20 +E+ E-- +++ ++
|H|+++ |
|+| +H+------- |
15 ++H+ ---+++ +H+------ ++
|N++H+-- +++--- +H+------++|
10 ++ +++ - +++ ---+H+ +++ +H+
| | +H+-----+H+------+H+-- |
5 ++| +++ ++
++N+N+--+N++ + + + + + |
0 ++---------+----------+---------+----------+----------+----------+---++
0 10 20 30 40 50 60
Number of threads
atomic_add-bench: 5000000 ops/thread, [0,128] range
160 ++---------+---------+----------+---------+----------+----------+---++
+ atomic +-E--+ + + + + + |
140 +cmpxchg +-H--+ +++ +++ ++
| master +-N--+ E--------E------+E+------++|
120 ++ --| | +++ E+
| -- +++ +++ ++|
100 ++ - ++
| +++- +++ ++|
80 ++ -+E+ -+H+------+H+------H--------++
| ---- ---- +++ H|
| ---+E+-----+E+- ---+H+ ++|
60 ++ +E+--- +++ ---+H+--- ++
| --+++ ---+H+-- |
40 ++ +E+-+H+--- ++
| +H+ |
20 +EE+ ++
+N+ + + + + + + |
0 ++N-N---N--+---------+----------+---------+----------+----------+---++
0 10 20 30 40 50 60
Number of threads
atomic_add-bench: 5000000 ops/thread, [0,1024] range
350 ++---------+---------+----------+---------+----------+----------+---++
+ atomic +-E--+ + + + + + |
300 +cmpxchg +-H--+ +++
| master +-N--+ +++ ||
| +++ | ----E|
250 ++ | ----E---- ++
| ----E--- | ---+H|
200 ++ -+E+--- +++ ---+H+--- ++
| ---- -+H+-- |
| +E+ +++ ---- +++ |
150 ++ ---+++ ---+H+- ++
| --- -+H+-- |
100 ++ ---+E+ ---- +++ ++
| +++ ---+E+-----+H+- |
| -+E+------+H+-- |
50 ++ +E+ ++
+EE+ + + + + + + |
0 ++N-N---N--+---------+----------+---------+----------+----------+---++
0 10 20 30 40 50 60
Number of threads
hi-res: http://imgur.com/a/fMRmq
For master I stopped measuring master after 8 threads, because there is little
point in measuring the well-known performance collapse of a contended lock.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1467054136-10430-21-git-send-email-cota@braap.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
---
target-i386/helper.h | 2 --
target-i386/mem_helper.c | 33 ---------------------------------
target-i386/translate.c | 15 ---------------
3 files changed, 50 deletions(-)
diff --git a/target-i386/helper.h b/target-i386/helper.h
index 729d4b6..4e859eb 100644
--- a/target-i386/helper.h
+++ b/target-i386/helper.h
@@ -1,8 +1,6 @@
DEF_HELPER_FLAGS_4(cc_compute_all, TCG_CALL_NO_RWG_SE, tl, tl, tl, tl, int)
DEF_HELPER_FLAGS_4(cc_compute_c, TCG_CALL_NO_RWG_SE, tl, tl, tl, tl, int)
-DEF_HELPER_0(lock, void)
-DEF_HELPER_0(unlock, void)
DEF_HELPER_3(write_eflags, void, env, tl, i32)
DEF_HELPER_1(read_eflags, tl, env)
DEF_HELPER_2(divb_AL, void, env, tl)
diff --git a/target-i386/mem_helper.c b/target-i386/mem_helper.c
index c4b5c5b..70f6766 100644
--- a/target-i386/mem_helper.c
+++ b/target-i386/mem_helper.c
@@ -25,39 +25,6 @@
#include "qemu/int128.h"
#include "tcg.h"
-/* broken thread support */
-
-#if defined(CONFIG_USER_ONLY)
-QemuMutex global_cpu_lock;
-
-void helper_lock(void)
-{
- qemu_mutex_lock(&global_cpu_lock);
-}
-
-void helper_unlock(void)
-{
- qemu_mutex_unlock(&global_cpu_lock);
-}
-
-void helper_lock_init(void)
-{
- qemu_mutex_init(&global_cpu_lock);
-}
-#else
-void helper_lock(void)
-{
-}
-
-void helper_unlock(void)
-{
-}
-
-void helper_lock_init(void)
-{
-}
-#endif
-
void helper_cmpxchg8b_unlocked(CPUX86State *env, target_ulong a0)
{
uintptr_t ra = GETPC();
diff --git a/target-i386/translate.c b/target-i386/translate.c
index c8827f3..ac3f6f4 100644
--- a/target-i386/translate.c
+++ b/target-i386/translate.c
@@ -4537,10 +4537,6 @@ static target_ulong disas_insn(CPUX86State *env, DisasContext *s,
s->aflag = aflag;
s->dflag = dflag;
- /* lock generation */
- if (prefixes & PREFIX_LOCK)
- gen_helper_lock();
-
/* now check op code */
reswitch:
switch(b) {
@@ -8211,20 +8207,11 @@ static target_ulong disas_insn(CPUX86State *env, DisasContext *s,
default:
goto unknown_op;
}
- /* lock generation */
- if (s->prefix & PREFIX_LOCK)
- gen_helper_unlock();
return s->pc;
illegal_op:
- if (s->prefix & PREFIX_LOCK)
- gen_helper_unlock();
- /* XXX: ensure that no lock was generated */
gen_illegal_opcode(s);
return s->pc;
unknown_op:
- if (s->prefix & PREFIX_LOCK)
- gen_helper_unlock();
- /* XXX: ensure that no lock was generated */
gen_unknown_opcode(env, s);
return s->pc;
}
@@ -8316,8 +8303,6 @@ void tcg_x86_init(void)
offsetof(CPUX86State, bnd_regs[i].ub),
bnd_regu_names[i]);
}
-
- helper_lock_init();
}
/* generate intermediate code for basic block 'tb'. */
--
2.7.4
next prev parent reply other threads:[~2016-10-22 21:05 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-22 21:04 [Qemu-devel] [PULL 00/35] cmpxchg atomic operations Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 01/35] atomics: add atomic_xor Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 02/35] atomics: add atomic_op_fetch variants Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 03/35] exec: Avoid direct references to Int128 parts Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 04/35] int128: Use __int128 if available Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 05/35] int128: Add int128_make128 Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 07/35] linux-user: enable parallel code generation on clone Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 08/35] cputlb: Replace SHIFT with DATA_SIZE Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 09/35] cputlb: Move probe_write out of softmmu_template.h Richard Henderson
2016-10-22 21:04 ` [Qemu-devel] [PULL 10/35] cputlb: Remove includes from softmmu_template.h Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 11/35] cputlb: Move most of iotlb code out of line Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 12/35] cputlb: Tidy some macros Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 13/35] tcg: Add atomic helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 14/35] tcg: Add atomic128 helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 15/35] tcg: Add CONFIG_ATOMIC64 Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 16/35] tcg: Emit barriers with parallel_cpus Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 17/35] target-i386: emulate LOCK'ed cmpxchg using cmpxchg helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 18/35] target-i386: emulate LOCK'ed OP instructions using atomic helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 19/35] target-i386: emulate LOCK'ed INC using atomic helper Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 20/35] target-i386: emulate LOCK'ed NOT " Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 21/35] target-i386: emulate LOCK'ed NEG using cmpxchg helper Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 22/35] target-i386: emulate LOCK'ed XADD using atomic helper Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 23/35] target-i386: emulate LOCK'ed BTX ops using atomic helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 24/35] target-i386: emulate XCHG using atomic helper Richard Henderson
2016-10-22 21:05 ` Richard Henderson [this message]
2016-10-22 21:05 ` [Qemu-devel] [PULL 26/35] tests: add atomic_add-bench Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 27/35] target-arm: Rearrange aa32 load and store functions Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 28/35] target-arm: emulate LL/SC using cmpxchg helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 29/35] target-arm: emulate SWP with atomic_xchg helper Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 30/35] target-arm: emulate aarch64's LL/SC using cmpxchg helpers Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 31/35] linux-user: remove handling of ARM's EXCP_STREX Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 32/35] linux-user: remove handling of aarch64's EXCP_STREX Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 33/35] target-arm: remove EXCP_STREX + cpu_exclusive_{test, info} Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 34/35] target-alpha: Introduce MMU_PHYS_IDX Richard Henderson
2016-10-22 21:05 ` [Qemu-devel] [PULL 35/35] target-alpha: Emulate LL/SC using cmpxchg helpers Richard Henderson
2016-10-24 10:51 ` [Qemu-devel] [PULL 00/35] cmpxchg atomic operations Peter Maydell
2016-10-24 12:37 ` Paolo Bonzini
2016-10-24 14:02 ` Peter Maydell
2016-10-24 17:27 ` Richard Henderson
2016-10-24 18:02 ` Peter Maydell
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1477170324-17783-26-git-send-email-rth@twiddle.net \
--to=rth@twiddle.net \
--cc=cota@braap.org \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).