public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
* [Bug rtl-optimization/21150] New: Suboptimal byte extraction from larger integers
@ 2005-04-21 13:10 vda at port dot imtp dot ilyichevsk dot odessa dot ua
2005-04-21 13:13 ` [Bug rtl-optimization/21150] " vda at port dot imtp dot ilyichevsk dot odessa dot ua
` (3 more replies)
0 siblings, 4 replies; 9+ messages in thread
From: vda at port dot imtp dot ilyichevsk dot odessa dot ua @ 2005-04-21 13:10 UTC (permalink / raw)
To: gcc-bugs
Bytes are typically extracted from e.g. u64's by something like
#define D5(v) (((v) >> 40) & 0xff)
Testcase shows that gcc does not optimize this "good enough".
--
Summary: Suboptimal byte extraction from larger integers
Product: gcc
Version: 3.4.3
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: rtl-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: vda at port dot imtp dot ilyichevsk dot odessa dot ua
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i386-pc-linux-gnu
GCC host triplet: i386-pc-linux-gnu
GCC target triplet: i386-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150
^ permalink raw reply [flat|nested] 9+ messages in thread* [Bug rtl-optimization/21150] Suboptimal byte extraction from larger integers 2005-04-21 13:10 [Bug rtl-optimization/21150] New: Suboptimal byte extraction from larger integers vda at port dot imtp dot ilyichevsk dot odessa dot ua @ 2005-04-21 13:13 ` vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-04-21 13:19 ` [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits pinskia at gcc dot gnu dot org ` (2 subsequent siblings) 3 siblings, 0 replies; 9+ messages in thread From: vda at port dot imtp dot ilyichevsk dot odessa dot ua @ 2005-04-21 13:13 UTC (permalink / raw) To: gcc-bugs ------- Additional Comments From vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-04-21 13:12 ------- Created an attachment (id=8701) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8701&action=view) generate assembly with -S and compare results -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 ^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits 2005-04-21 13:10 [Bug rtl-optimization/21150] New: Suboptimal byte extraction from larger integers vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-04-21 13:13 ` [Bug rtl-optimization/21150] " vda at port dot imtp dot ilyichevsk dot odessa dot ua @ 2005-04-21 13:19 ` pinskia at gcc dot gnu dot org 2005-04-24 13:26 ` vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-07-04 22:21 ` pinskia at gcc dot gnu dot org 3 siblings, 0 replies; 9+ messages in thread From: pinskia at gcc dot gnu dot org @ 2005-04-21 13:19 UTC (permalink / raw) To: gcc-bugs -- What |Removed |Added ---------------------------------------------------------------------------- Keywords| |missed-optimization Summary|Suboptimal byte extraction |Suboptimal byte extraction |from larger integers |from 64bits http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 ^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits 2005-04-21 13:10 [Bug rtl-optimization/21150] New: Suboptimal byte extraction from larger integers vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-04-21 13:13 ` [Bug rtl-optimization/21150] " vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-04-21 13:19 ` [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits pinskia at gcc dot gnu dot org @ 2005-04-24 13:26 ` vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-07-04 22:21 ` pinskia at gcc dot gnu dot org 3 siblings, 0 replies; 9+ messages in thread From: vda at port dot imtp dot ilyichevsk dot odessa dot ua @ 2005-04-24 13:26 UTC (permalink / raw) To: gcc-bugs ------- Additional Comments From vda at port dot imtp dot ilyichevsk dot odessa dot ua 2005-04-24 13:26 ------- I don't think that bug description is correct. I believe similar observation will be valid for byte extraction from u32 and u16, and for u16-from-u32, etc. Update for latest gcc. This is what 4.0.0 produces from the testcase: # gcc -O2 -fomit-frame-pointer -S helper.c # cat helper.s [I removed non-essential stuff] a: movl v+8, %eax shrl $8, %eax xorb v, %al xorb v+18, %al xorb v+27, %al xorb v+36, %al movl v+40, %edx movl v+44, %ecx movl %ecx, %edx xorl %ecx, %ecx shrl $8, %edx xorl %edx, %eax xorb v+54, %al xorb v+63, %al movzbl %al, %eax ret b: movl v+8, %eax movl v+12, %edx shrdl $8, %edx, %eax shrl $8, %edx xorb v, %al movl v+16, %edx movl v+20, %ecx shrdl $16, %ecx, %edx shrl $16, %ecx xorl %edx, %eax movl v+24, %edx movl v+28, %ecx shrdl $24, %ecx, %edx shrl $24, %ecx xorl %edx, %eax xorb v+36, %al movl v+40, %edx movl v+44, %ecx movl %ecx, %edx xorl %ecx, %ecx shrl $8, %edx xorl %edx, %eax xorb v+54, %al xorb v+63, %al movzbl %al, %eax ret c: movb v+9, %al xorb v, %al xorb v+18, %al xorb v+27, %al xorb v+36, %al xorb v+45, %al xorb v+54, %al xorb v+63, %al movzbl %al, %eax ret d: movl v+8, %eax movl v+12, %edx shrdl $8, %edx, %eax shrl $8, %edx xorb v, %al movl v+16, %edx movl v+20, %ecx shrdl $16, %ecx, %edx shrl $16, %ecx xorl %edx, %eax movl v+24, %edx movl v+28, %ecx shrdl $24, %ecx, %edx shrl $24, %ecx xorl %edx, %eax xorb v+36, %al movl v+40, %edx movl v+44, %ecx movl %ecx, %edx xorl %ecx, %ecx shrl $8, %edx xorl %edx, %eax xorb v+54, %al xorb v+63, %al movzbl %al, %eax ret As you can see, a,b and d results are far from optimal, while c is almost perfect. Note that people typically use d, i.e. this: #define D7(v) (((v) >> 56)) #define D6(v) (((v) >> 48) & 0xff) #define D5(v) (((v) >> 40) & 0xff) #define D4(v) (((v) >> 32) & 0xff) #define D3(v) (((v) >> 24) & 0xff) #define D2(v) (((v) >> 16) & 0xff) #define D1(v) (((v) >> 8) & 0xff) #define D0(v) ((v) & 0xff) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 ^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits 2005-04-21 13:10 [Bug rtl-optimization/21150] New: Suboptimal byte extraction from larger integers vda at port dot imtp dot ilyichevsk dot odessa dot ua ` (2 preceding siblings ...) 2005-04-24 13:26 ` vda at port dot imtp dot ilyichevsk dot odessa dot ua @ 2005-07-04 22:21 ` pinskia at gcc dot gnu dot org 3 siblings, 0 replies; 9+ messages in thread From: pinskia at gcc dot gnu dot org @ 2005-07-04 22:21 UTC (permalink / raw) To: gcc-bugs ------- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-04 22:21 ------- Confirmed, this is most likely a sub register problem. -- What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed| |1 Last reconfirmed|0000-00-00 00:00:00 |2005-07-04 22:21:33 date| | http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 ^ permalink raw reply [flat|nested] 9+ messages in thread
[parent not found: <bug-21150-1101@http.gcc.gnu.org/bugzilla/>]
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits [not found] <bug-21150-1101@http.gcc.gnu.org/bugzilla/> @ 2007-11-09 19:48 ` rask at gcc dot gnu dot org 2008-09-14 3:54 ` pinskia at gcc dot gnu dot org 1 sibling, 0 replies; 9+ messages in thread From: rask at gcc dot gnu dot org @ 2007-11-09 19:48 UTC (permalink / raw) To: gcc-bugs ------- Comment #4 from rask at gcc dot gnu dot org 2007-11-09 19:48 ------- I think this might be a middle-end issue related to PR 7061 or PR 15184. We're doing slightly better with GCC 4.3.0 (because of subreg lowering, I guess), but not much (asm output with -dp for readability): a: movl v+44, %eax # 53 *movsi_1/1 [length = 5] movl v+8, %edx # 23 *movsi_1/1 [length = 6] shrl $8, %eax # 54 *lshrsi3_1/1 [length = 3] xorb v+36, %al # 11 *xorqi_1/1 [length = 6] xorb v, %al # 13 *xorqi_1/1 [length = 6] xorb v+54, %al # 17 *xorqi_1/1 [length = 6] xorb v+63, %al # 21 *xorqi_1/1 [length = 6] shrl $8, %edx # 24 *lshrsi3_1/1 [length = 3] xorl %edx, %eax # 66 *xorsi_1/1 [length = 2] xorb v+18, %al # 29 *xorqi_1/1 [length = 6] xorb v+27, %al # 33 *xorqi_1/1 [length = 6] ret # 69 return_internal [length = 1] b: pushl %ebx # 75 *pushsi2 [length = 1] movl v+20, %edx # 69 *movsi_1/1 [length = 6] movl v+12, %ebx # 67 *movsi_1/1 [length = 6] movl v+8, %ecx # 66 *movsi_1/1 [length = 6] movl v+16, %eax # 68 *movsi_1/1 [length = 5] shrdl $8, %ebx, %ecx # 81 x86_shrd_1/1 [length = 4] shrdl $16, %edx, %eax # 83 x86_shrd_1/1 [length = 4] movl v+24, %edx # 71 *movsi_1/1 [length = 6] xorl %ecx, %eax # 70 *xorsi_1/1 [length = 2] movl v+28, %ecx # 72 *movsi_1/1 [length = 6] xorb v, %al # 13 *xorqi_1/1 [length = 6] popl %ebx # 78 popsi1 [length = 1] shrdl $24, %ecx, %edx # 85 x86_shrd_1/1 [length = 4] xorl %edx, %eax # 73 *xorsi_1/1 [length = 2] movl v+44, %edx # 53 *movsi_1/1 [length = 6] xorb v+36, %al # 21 *xorqi_1/1 [length = 6] shrl $8, %edx # 54 *lshrsi3_1/1 [length = 3] xorl %edx, %eax # 74 *xorsi_1/1 [length = 2] xorb v+54, %al # 29 *xorqi_1/1 [length = 6] xorb v+63, %al # 33 *xorqi_1/1 [length = 6] ret # 79 return_internal [length = 1] c: movzbl v+9, %eax # 7 *movqi_1/3 [length = 7] xorb v+18, %al # 8 *xorqi_1/1 [length = 6] xorb v, %al # 9 *xorqi_1/1 [length = 6] xorb v+27, %al # 10 *xorqi_1/1 [length = 6] xorb v+36, %al # 11 *xorqi_1/1 [length = 6] xorb v+45, %al # 12 *xorqi_1/1 [length = 6] xorb v+54, %al # 13 *xorqi_1/1 [length = 6] xorb v+63, %al # 14 *xorqi_1/1 [length = 6] ret # 33 return_internal [length = 1] d: pushl %ebx # 75 *pushsi2 [length = 1] movl v+20, %edx # 69 *movsi_1/1 [length = 6] movl v+12, %ebx # 67 *movsi_1/1 [length = 6] movl v+8, %ecx # 66 *movsi_1/1 [length = 6] movl v+16, %eax # 68 *movsi_1/1 [length = 5] shrdl $8, %ebx, %ecx # 81 x86_shrd_1/1 [length = 4] shrdl $16, %edx, %eax # 83 x86_shrd_1/1 [length = 4] movl v+24, %edx # 71 *movsi_1/1 [length = 6] xorl %ecx, %eax # 70 *xorsi_1/1 [length = 2] movl v+28, %ecx # 72 *movsi_1/1 [length = 6] xorb v, %al # 13 *xorqi_1/1 [length = 6] popl %ebx # 78 popsi1 [length = 1] shrdl $24, %ecx, %edx # 85 x86_shrd_1/1 [length = 4] xorl %edx, %eax # 73 *xorsi_1/1 [length = 2] movl v+44, %edx # 53 *movsi_1/1 [length = 6] xorb v+36, %al # 21 *xorqi_1/1 [length = 6] shrl $8, %edx # 54 *lshrsi3_1/1 [length = 3] xorl %edx, %eax # 74 *xorsi_1/1 [length = 2] xorb v+54, %al # 29 *xorqi_1/1 [length = 6] xorb v+63, %al # 33 *xorqi_1/1 [length = 6] ret # 79 return_internal [length = 1] .ident "GCC: (GNU) 4.3.0 20071102 (experimental)" -- rask at gcc dot gnu dot org changed: What |Removed |Added ---------------------------------------------------------------------------- Known to fail| |4.3.0 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 ^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits [not found] <bug-21150-1101@http.gcc.gnu.org/bugzilla/> 2007-11-09 19:48 ` rask at gcc dot gnu dot org @ 2008-09-14 3:54 ` pinskia at gcc dot gnu dot org 1 sibling, 0 replies; 9+ messages in thread From: pinskia at gcc dot gnu dot org @ 2008-09-14 3:54 UTC (permalink / raw) To: gcc-bugs ------- Comment #5 from pinskia at gcc dot gnu dot org 2008-09-14 03:53 ------- IRA is slightly worse, it uses one extra register .... -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 ^ permalink raw reply [flat|nested] 9+ messages in thread
[parent not found: <bug-21150-4@http.gcc.gnu.org/bugzilla/>]
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits [not found] <bug-21150-4@http.gcc.gnu.org/bugzilla/> @ 2013-01-18 11:12 ` vda.linux at googlemail dot com 2021-07-25 0:51 ` pinskia at gcc dot gnu.org 1 sibling, 0 replies; 9+ messages in thread From: vda.linux at googlemail dot com @ 2013-01-18 11:12 UTC (permalink / raw) To: gcc-bugs http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 Denis Vlasenko <vda.linux at googlemail dot com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |vda.linux at googlemail dot | |com --- Comment #6 from Denis Vlasenko <vda.linux at googlemail dot com> 2013-01-18 11:12:18 UTC --- Guess this can be closed now. All four cases look good: $ cat helper-4.6.3-O2.asm helper-4.6.3-O2.o: file format elf32-i386 ... 00000000 <a>: 0: 0f b6 05 2d 00 00 00 movzbl 0x2d,%eax 7: 32 05 24 00 00 00 xor 0x24,%al d: 32 05 00 00 00 00 xor 0x0,%al 13: 32 05 36 00 00 00 xor 0x36,%al 19: 32 05 3f 00 00 00 xor 0x3f,%al 1f: 32 05 09 00 00 00 xor 0x9,%al 25: 32 05 12 00 00 00 xor 0x12,%al 2b: 32 05 1b 00 00 00 xor 0x1b,%al 31: c3 ret Disassembly of section .text.b: 00000000 <b>: 0: 0f b6 05 12 00 00 00 movzbl 0x12,%eax 7: 32 05 09 00 00 00 xor 0x9,%al d: 32 05 00 00 00 00 xor 0x0,%al 13: 32 05 1b 00 00 00 xor 0x1b,%al 19: 32 05 24 00 00 00 xor 0x24,%al 1f: 32 05 2d 00 00 00 xor 0x2d,%al 25: 32 05 36 00 00 00 xor 0x36,%al 2b: 32 05 3f 00 00 00 xor 0x3f,%al 31: c3 ret Disassembly of section .text.c: 00000000 <c>: 0: 0f b6 05 09 00 00 00 movzbl 0x9,%eax 7: 32 05 00 00 00 00 xor 0x0,%al d: 32 05 12 00 00 00 xor 0x12,%al 13: 32 05 1b 00 00 00 xor 0x1b,%al 19: 32 05 24 00 00 00 xor 0x24,%al 1f: 32 05 2d 00 00 00 xor 0x2d,%al 25: 32 05 36 00 00 00 xor 0x36,%al 2b: 32 05 3f 00 00 00 xor 0x3f,%al 31: c3 ret Disassembly of section .text.d: 00000000 <d>: 0: 0f b6 05 12 00 00 00 movzbl 0x12,%eax 7: 32 05 09 00 00 00 xor 0x9,%al d: 32 05 00 00 00 00 xor 0x0,%al 13: 32 05 1b 00 00 00 xor 0x1b,%al 19: 32 05 24 00 00 00 xor 0x24,%al 1f: 32 05 2d 00 00 00 xor 0x2d,%al 25: 32 05 36 00 00 00 xor 0x36,%al 2b: 32 05 3f 00 00 00 xor 0x3f,%al 31: c3 ret Curiously, -Os manages to squeeze two more bytes out of it. helper-4.6.3-Os.o: file format elf32-i386 00000000 <a>: 0: a0 2d 00 00 00 mov 0x2d,%al ^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^ better than movzbl 5: 33 05 24 00 00 00 xor 0x24,%eax << why %eax? oh well... b: 33 05 00 00 00 00 xor 0x0,%eax 11: 32 05 36 00 00 00 xor 0x36,%al 17: 32 05 3f 00 00 00 xor 0x3f,%al 1d: 32 05 09 00 00 00 xor 0x9,%al 23: 32 05 12 00 00 00 xor 0x12,%al 29: 32 05 1b 00 00 00 xor 0x1b,%al 2f: c3 ret ^ permalink raw reply [flat|nested] 9+ messages in thread
* [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits [not found] <bug-21150-4@http.gcc.gnu.org/bugzilla/> 2013-01-18 11:12 ` vda.linux at googlemail dot com @ 2021-07-25 0:51 ` pinskia at gcc dot gnu.org 1 sibling, 0 replies; 9+ messages in thread From: pinskia at gcc dot gnu.org @ 2021-07-25 0:51 UTC (permalink / raw) To: gcc-bugs https://gcc.gnu.org/bugzilla/show_bug.cgi?id=21150 Andrew Pinski <pinskia at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Target Milestone|--- |4.6.3 Resolution|--- |FIXED --- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> --- Fixed in GCC 4.6.3 and above. ^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2021-07-25 0:51 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-04-21 13:10 [Bug rtl-optimization/21150] New: Suboptimal byte extraction from larger integers vda at port dot imtp dot ilyichevsk dot odessa dot ua
2005-04-21 13:13 ` [Bug rtl-optimization/21150] " vda at port dot imtp dot ilyichevsk dot odessa dot ua
2005-04-21 13:19 ` [Bug rtl-optimization/21150] Suboptimal byte extraction from 64bits pinskia at gcc dot gnu dot org
2005-04-24 13:26 ` vda at port dot imtp dot ilyichevsk dot odessa dot ua
2005-07-04 22:21 ` pinskia at gcc dot gnu dot org
[not found] <bug-21150-1101@http.gcc.gnu.org/bugzilla/>
2007-11-09 19:48 ` rask at gcc dot gnu dot org
2008-09-14 3:54 ` pinskia at gcc dot gnu dot org
[not found] <bug-21150-4@http.gcc.gnu.org/bugzilla/>
2013-01-18 11:12 ` vda.linux at googlemail dot com
2021-07-25 0:51 ` pinskia at gcc dot gnu.org
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 read-only IMAP folder(s) and NNTP newsgroup(s).