Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
MG == Michał Górny mgo...@gentoo.org writes: MG sse4_1 - Enable SSE4.1 instruction support MG sse4_2 - Enable SSE4.2 instruction support Does everything use the _ versions? The docs should cover the . versions, as well, if anything still expects them. I've had to use sse4.1 and sse4.2, too, for some stuff. (Obviously . xor _ in the ebuilds (a/o eclasses?) would be nice.) -JimC -- James Cloos cl...@jhcloos.com OpenPGP: 0x997A9F17ED7DAEA6
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Tue, 20 Jan 2015 09:52:40 +0100 Michał Górny mgo...@gentoo.org wrote: Dnia 2015-01-19, o godz. 11:38:26 Alexis Ballier aball...@gentoo.org napisał(a): On Sun, 18 Jan 2015 21:44:05 +0100 Michał Górny mgo...@gentoo.org wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set e... are you aware that these descriptions are close to useless ? 'foo - enable foo' - thanks for the information I couldn't have guessed... It's not my fault we have to put some description there! I know; I'm just ranting because you could grep use.local.desc to get more informative ones for some of them; then we can start discussing on how to improve the other ones :) Alexis.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
Dnia 2015-01-19, o godz. 11:38:26 Alexis Ballier aball...@gentoo.org napisał(a): On Sun, 18 Jan 2015 21:44:05 +0100 Michał Górny mgo...@gentoo.org wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set e... are you aware that these descriptions are close to useless ? 'foo - enable foo' - thanks for the information I couldn't have guessed... It's not my fault we have to put some description there! you already have more useful ones available: flag name=fma3Enables FMA3 optimizations: AMD processors starting with Piledriver architecture and Intel Haswell based processors or later./flag flag name=fma4Enables FMA4 optimizations: AMD processors starting with Bulldozer architecture./flag flag name=sse4_2Enables SSE4.2 optimizations: Nehalem-based Intel Core i7 or later./flag flag name=ssse3Faster floating point optimization for SSSE3 capable chips (Intel Core 2 and later chips)/flag and so on... Feel free to update them yourself. As far as I'm concerned, those descriptions are worse than useless because they end up being confusing. Many CPU lines had weird models with more or less instruction sets, so brand name aren't really helpful. It's better to focus on /proc/cpuinfo. -- Best regards, Michał Górny pgpwEywupN5yb.pgp Description: OpenPGP digital signature
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Mon, Jan 19, 2015 at 4:41 AM, Alexis Ballier aball...@gentoo.org wrote: On Sun, 18 Jan 2015 15:15:22 -0800 Matt Turner matts...@gentoo.org wrote: mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support We shouldn't have an sse4 USE flag. It's either one of the three below, but SSE4 by itself isn't a thing. wikipedia says it is sse4.1 + sse4.2 Some CPUs don't have 4.2, and some don't have 4a. You'd still need the separate flags, so an extraneous combined flag is a bit silly.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Mon, 19 Jan 2015 08:13:46 +0800 Patrick Lauer wrote: On Sunday 18 January 2015 21:44:05 Michał Górny wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set Those are kinda mostly dead (no new CPUs have them anymore) 1) Modern AMD CPUs still support them (and can give performance benefits from using them. 2) I'm writing right now from an old host with CPU supporting them. aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set Not sure if this one needs to be specified - amd64 always has it, and on x86 your code should do cpu feature detection anyway 1) No. We have cpudetection USE flag if user wants it. 2) Not all VIA CPUs support MMX. 3) Run-time CPU detection is *always slower* than compiled in code. (Learn the difference between using GOTs or long jupms and not using them, especiall on non-amd64 sysmems.) mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions Kinda very dead, even more than 3dnow So what? There are still people using it. popcnt - Enable popcnt instruction support Why?! Why not? It is an instruction set outside from official SSE4.2 standard, as well as LZCNT by the way. sse - Use the SSE instruction set Always exists on amd64, so this would be x86 special sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set Wow, such a wonderful mess :) Very productive comment... So half of those are obsolete/dead, and the other half you need to do proper feature detection - why do we want that as useflags again? Because we have users interested in these flags, including myself but not limited to. Best regards, Andrew Savchenko pgplC2PtOBYLx.pgp Description: PGP signature
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, 18 Jan 2015 15:15:22 -0800 Matt Turner matts...@gentoo.org wrote: mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support We shouldn't have an sse4 USE flag. It's either one of the three below, but SSE4 by itself isn't a thing. wikipedia says it is sse4.1 + sse4.2 Alexis.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, 18 Jan 2015 21:44:05 +0100 Michał Górny mgo...@gentoo.org wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set e... are you aware that these descriptions are close to useless ? 'foo - enable foo' - thanks for the information I couldn't have guessed... you already have more useful ones available: flag name=fma3Enables FMA3 optimizations: AMD processors starting with Piledriver architecture and Intel Haswell based processors or later./flag flag name=fma4Enables FMA4 optimizations: AMD processors starting with Bulldozer architecture./flag flag name=sse4_2Enables SSE4.2 optimizations: Nehalem-based Intel Core i7 or later./flag flag name=ssse3Faster floating point optimization for SSSE3 capable chips (Intel Core 2 and later chips)/flag and so on... Alexis.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Mon, 19 Jan 2015 08:13:46 +0800 Patrick Lauer patr...@gentoo.org wrote: So half of those are obsolete/dead, and the other half you need to do proper feature detection - why do we want that as useflags again? http://article.gmane.org/gmane.linux.gentoo.devel/94299
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, Jan 18, 2015 at 6:13 PM, Patrick Lauer patr...@gentoo.org wrote: On Sunday 18 January 2015 21:44:05 Michał Górny wrote: popcnt - Enable popcnt instruction support Why?! Because Intel and AMD support it via different cpuinfo feature names. It is popcnt on Intel, and abm on AMD. The description of the flag should also mention that it is included in feature abm on AMD CPUs (and Intel CPUs, but since they list popcnt separately, too, that might be confusing).
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, 18 Jan 2015 20:28:27 -0600 Gordon Pettey petteyg...@gmail.com wrote: Because Intel and AMD support it via different cpuinfo feature names. It is popcnt on Intel, and abm on AMD. The description of the flag should also mention that it is included in feature abm on AMD CPUs (and Intel CPUs, but since they list popcnt separately, too, that might be confusing). That's probably not quite correct. I have both of them in /proc/cpuinfo on AMD Phenom II CPU.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, 18 Jan 2015 21:44:05 +0100 Michał Górny mgo...@gentoo.org wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set There are other cool instructions on modern CPUs too. For example, iommu. There are versions of them in kernel config options. Maybe genkernel should depend on them? IDK, because I configure kernel config by hand.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Mon, 19 Jan 2015 08:13:46 +0800 Patrick Lauer patr...@gentoo.org wrote: 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set Those are kinda mostly dead (no new CPUs have them anymore) My CPU still has them. mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions Kinda very dead, even more than 3dnow I still have CPU with this instruction too.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, Jan 18, 2015 at 7:13 PM, Patrick Lauer patr...@gentoo.org wrote: 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set Those are kinda mostly dead (no new CPUs have them anymore) I'm sure there are still plenty of these out in the wild. I know I have one. So half of those are obsolete/dead, and the other half you need to do proper feature detection - why do we want that as useflags again? The alternative would be to hard-disable them for the relevant packages where runtime detection is not implemented (probably rare).
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
Dnia 2015-01-18, o godz. 15:15:22 Matt Turner matts...@gentoo.org napisał(a): On Sun, Jan 18, 2015 at 12:44 PM, Michał Górny mgo...@gentoo.org wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set There are fma3 and fma4 instruction sets. We need to determine which one the packages that have this flag are actually using and name the flag appropriately. Ok then: fma3 - Use the Fused Multiply Add v3 instruction set fma4 - Use the Fused Multiply Add v4 instruction set mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support We shouldn't have an sse4 USE flag. It's either one of the three below, but SSE4 by itself isn't a thing. Ok. dev-libs/marisa has --enable for all *four* of them :). And some totally crazy, broken --enable magic which looks very mismatched... sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set -- Best regards, Michał Górny pgpoRXAFBBcIa.pgp Description: OpenPGP digital signature
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sun, Jan 18, 2015 at 3:23 PM, Michał Górny mgo...@gentoo.org wrote: Dnia 2015-01-18, o godz. 15:15:22 Matt Turner matts...@gentoo.org napisał(a): On Sun, Jan 18, 2015 at 12:44 PM, Michał Górny mgo...@gentoo.org wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set There are fma3 and fma4 instruction sets. We need to determine which one the packages that have this flag are actually using and name the flag appropriately. Ok then: fma3 - Use the Fused Multiply Add v3 instruction set fma4 - Use the Fused Multiply Add v4 instruction set They're not actually versions 3 and 4, like SSE. The number is the number of arguments the instructions take. I'd just remove the 'v' from the description and call it good. See http://en.wikipedia.org/wiki/FMA_instruction_set for more information.
Re: [gentoo-dev] Review: desc/cpu_flags_x86.desc
On Sunday 18 January 2015 21:44:05 Michał Górny wrote: Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set Those are kinda mostly dead (no new CPUs have them anymore) aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set Not sure if this one needs to be specified - amd64 always has it, and on x86 your code should do cpu feature detection anyway mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions Kinda very dead, even more than 3dnow popcnt - Enable popcnt instruction support Why?! sse - Use the SSE instruction set Always exists on amd64, so this would be x86 special sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set Wow, such a wonderful mess :) So half of those are obsolete/dead, and the other half you need to do proper feature detection - why do we want that as useflags again?
[gentoo-dev] Review: desc/cpu_flags_x86.desc
Hello, I would like to commit the following flags as cpu_flags_x86_desc. The list combines global USE flags with some local USE flags I've been able to find. 3dnow - Use the 3DNow! instruction set 3dnowext - Use the Enhanced 3DNow! instruction set aes-ni - Enable support for Intel's AES instruction set (aes in cpuinfo) avx - Adds support for Advanced Vector Extensions instructions avx2 - Adds support for Advanced Vector Extensions 2 instructions fma - Use the Fused Multiply Add instruction set mmx - Use the MMX instruction set mmxext - Use the Extended MMX instruction set (intersection of Enhanced 3DNow! and SSE instruction sets) (3dnowext or sse in cpuinfo) padlock - Use VIA padlock instructions popcnt - Enable popcnt instruction support sse - Use the SSE instruction set sse2 - Use the SSE2 instruction set sse3 - Use the SSE3 instruction set (pni in cpuinfo) sse4 - Enable SSE4 instruction support sse4_1 - Enable SSE4.1 instruction support sse4_2 - Enable SSE4.2 instruction support sse4a - Enable SSE4a instruction support ssse3 - Use the SSSE3 instruction set -- Best regards, Michał Górny pgpbuKGc5cgoI.pgp Description: OpenPGP digital signature
