Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
AMD Zen/Ryzen thread
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 15, 16, 17 ... 20, 21, 22  Next  
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3367
Location: Gainesville, Florida

PostPosted: Sat May 20, 2017 3:24 pm    Post subject: Reply with quote

I have sysrescueCD-5.0.0 from 4/23.17 with kernel-4.9.24 and gcc-5.4.0. When I boot it I'll check and see if they make /dev/shm a symbolic link to /run/shm/ and proceed accordingly.

I was considering just using my Ubuntu based Linux Mint-kde version on sda1, install the required chroot packages on it, and us it for the new Gentoo install, following the outline I saw here:

http://www.wikihow.com/Install-Gentoo-Linux-from-Ubuntu I think it's from 2013, but once chrooted I'll pick up at the amd64 handbook right after chroot at "Configuring Portage" and proceed from there.

Anyone ever install from another distro, or is using the cd better?
_________________
Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.32, gcc-10.2.0, kernel-5.9.1-gentoo, 5.10.0-rc1 USE=experimental
Back to top
View user's profile Send private message
ct85711
Veteran
Veteran


Joined: 27 Sep 2005
Posts: 1791

PostPosted: Sat May 20, 2017 5:58 pm    Post subject: Reply with quote

I know I have used Ubuntu and Fedora's live cd's a couple times to install Gentoo without any difference. Overall, the process works the same way (only need to adjust as necessary on the mounting the stuff for the chroot) compared to the Gentoo's minimal cd. Though being able to install Gentoo from a terminal window in a GUI is much nicer than working from straight command line. Couple that to also having a browser open to follow the handbook and of course being able to idly surf the web while you wait for the compile to finish is nice. Personally, working from another distro would probably be easier in that everything is faster to load (such as loading a browser window); beyond that I don't think there is much difference.
Back to top
View user's profile Send private message
Chiitoo
Administrator
Administrator


Joined: 28 Feb 2010
Posts: 2155
Location: Here and Away Again

PostPosted: Mon May 22, 2017 12:17 pm    Post subject: ><)))°€ Reply with quote

NeddySeagoon wrote:
WINE used to be 32 bit only too. I don't know if its even possible to fix that and still support 32 bit Windows apps.

It's very much possible to build Wine as 64-bit only, but aside from a very specific 64-bit only purpose, it would not be very useful. :]

Generally a WoW64 Wine prefix is the most useful, or indeed a 32-bit only (some applications still have issues with a 64-bit enabled prefix, even if it supports 32 bits too). The build can still support both.
_________________
Kind regards,
Chiitoo.

You might remember me from Gentoo projects such as Forums, LXQt, Qt, and Wine.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 4267
Location: Illinois, USA

PostPosted: Mon May 22, 2017 2:14 pm    Post subject: Re: ><)))°€ Reply with quote

Chiitoo wrote:
Generally a WoW64 Wine prefix is the most useful, or indeed a 32-bit only (some applications still have issues with a 64-bit enabled prefix, even if it supports 32 bits too). The build can still support both.

How can I build a 32-bit only Wine?
Back to top
View user's profile Send private message
Chiitoo
Administrator
Administrator


Joined: 28 Feb 2010
Posts: 2155
Location: Here and Away Again

PostPosted: Tue May 23, 2017 9:08 am    Post subject: Re: ><)))°€ Reply with quote

Tony0945 wrote:
How can I build a 32-bit only Wine?

While I was mainly talking about creating a 32-bit prefix using the 32/64-bit build (via the WINEARCH variable for example), that is a good question, considering the new ebuilds seem to force USE="abi_x32_64" on amd64.

I'll see if we need to add a '-abi_x86_64' in '/usr/portage/profiles/arch/amd64/package.use.force'.
_________________
Kind regards,
Chiitoo.

You might remember me from Gentoo projects such as Forums, LXQt, Qt, and Wine.
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 4267
Location: Illinois, USA

PostPosted: Tue May 23, 2017 5:11 pm    Post subject: Re: ><)))°€ Reply with quote

Chiitoo wrote:
I'll see if we need to add a '-abi_x86_64' in '/usr/portage/profiles/arch/amd64/package.use.force'.


I checked and it's already there.
Back to top
View user's profile Send private message
Chiitoo
Administrator
Administrator


Joined: 28 Feb 2010
Posts: 2155
Location: Here and Away Again

PostPosted: Wed May 24, 2017 10:16 am    Post subject: Re: ><)))°€ Reply with quote

Tony0945 wrote:
Chiitoo wrote:
I'll see if we need to add a '-abi_x86_64' in '/usr/portage/profiles/arch/amd64/package.use.force'.


I checked and it's already there.

For the old/current 'app-emulation/wine', yeah, it's been there.

As of 2017-05-23 19:09:43 -0400, it's there for the new variants as well, so it should be possible to build without 64-bit support with just ABI_X86="32" (set via the USE-flag, or otherwise).

(A little tempted to split these Wine bits into the 'wine: questions on recent changes' topic, but eh.)


Meh. Getting more and more antsy here to get my Ryzen already!

I think I'll wait just a little bit longer, though it isn't easy.
_________________
Kind regards,
Chiitoo.

You might remember me from Gentoo projects such as Forums, LXQt, Qt, and Wine.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3367
Location: Gainesville, Florida

PostPosted: Thu May 25, 2017 12:50 pm    Post subject: Reply with quote

Anyone have any comments on gcc-6.3.0 problems?

On a fresh install on sda6 I got Gentoo installed, and booting but still can't get X up and functioning (that may be my problem, not gcc's (?).

Anyway, I can't get through even an emerge -e @system (or -e @world), either chrooted in, or booted to a terminal. Doesn't seem to matter if I try rebuilds with amd64 or ~amd64, or various reccommended make.conf settings. With all the things I have tried over days, I still have compiling and conflict problems especially with python, ruby, and pestering openssl/openssh BINDLIST or -BINDLIST (or no bindlist at all in make.conf). It's a mess of paradoxical problems that won't relent (tried many of the recommended fixes, none worked), of course preventing doing any rebuilds.

I then saw the" GCC 6.3 is not presently optimized for Ryzen" warning on https://wiki.gentoo.org/wiki/Ryzen#Kernel and thought maybe this is the main problem I have, and am considering reverting to gcc-5.4 with -march=bdver4 and the extra zen flags shown on the same ryzen wiki page.

I guess my basic questions are is anyone successfuly running gcc-6 with ryzen (stable or ~arch) with no problems as described above, or gone back to gcc-5.4, rebuilt, and resolved any issues experienced with gcc-6? In other words, are my problems self-inflicted, or actually gcc-6.3 related? Would gcc-5.4 or even moving to gcc-7.1 really be a better move, or not?

Any advice or tips is greatly appreciated.

One other thing. I have another Gentoo (AM3+ system) on sda11 I need to do, and was wondering if there are problems with just chrooting into it, editing make.conf, and rebuiding world with the bdver4 flags, and perhaps building a new 4.11.2 kernel? Or, is that just wishful thinking? :roll:
_________________
Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.32, gcc-10.2.0, kernel-5.9.1-gentoo, 5.10.0-rc1 USE=experimental
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 46684
Location: 56N 3W

PostPosted: Thu May 25, 2017 1:12 pm    Post subject: Reply with quote

wrc1944,

Stop trying to eat an elephant all in one go.

Build problems may be the Ryzen hardware.
Conflict problems are certainly not.

Pick one particular plateful of your elephant and we can fix that before moving on to the next meal.

Its not trivial to move from ~arch to arch. If you manage to downgrade glibc you will be in a position
to learn things about Gentoo you can learn no other way.
The simile to having a tiger by the tail is deliberate.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3367
Location: Gainesville, Florida

PostPosted: Thu May 25, 2017 1:41 pm    Post subject: Reply with quote

NeddySeagoon wrote:
Quote:
Its not trivial to move from ~arch to arch. If you manage to downgrade glibc you will be in a position
to learn things about Gentoo you can learn no other way.

Oh yeah.... Forgot about the little downgrading glibc detail. :oops: Point taken.

But going to gcc-5.4 and bdver4 flags would be OK on ~arch, correct? Guess the point would be will it make any real difference? That gcc-6.3.0 warning has me worried.
_________________
Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.32, gcc-10.2.0, kernel-5.9.1-gentoo, 5.10.0-rc1 USE=experimental
Back to top
View user's profile Send private message
Tony0945
Advocate
Advocate


Joined: 25 Jul 2006
Posts: 4267
Location: Illinois, USA

PostPosted: Thu May 25, 2017 2:50 pm    Post subject: Reply with quote

wrc1944 wrote:
But going to gcc-5.4 and bdver4 flags would be OK on ~arch, correct? Guess the point would be will it make any real difference? That gcc-6.3.0 warning has me worried.


Wasn't there a thread with a link that said bulldozer instructions were not built into Zen and will make it crash?

Re gcc-6.3.0:

I did it on (k10) by the old-fashioned, "you don't really have to do this" way.

Code:
emerge -1v gcc
emerge -1v libtool
emerge -1v glibc
emerge -e @system
emerge -e @world
I have --keep-going in my make,conf and had to do "emerge --resume" several times, but eventually they all compiled, which tells me that something is wrong with dependencies.

On one machine I used a script from khayyam that emerge -e world --excluding things in system, but I don't have it anymore. It's somewhere in this forum.
Straight out "emerge -e @world" took 11 or 12 hours on a Phenom II X6 8G memory, 10,000 RPM drive.[/quote]
Back to top
View user's profile Send private message
Naib
Watchman
Watchman


Joined: 21 May 2004
Posts: 5855
Location: Removed by Neddy

PostPosted: Thu May 25, 2017 4:46 pm    Post subject: Reply with quote

Tony0945 wrote:
wrc1944 wrote:
But going to gcc-5.4 and bdver4 flags would be OK on ~arch, correct? Guess the point would be will it make any real difference? That gcc-6.3.0 warning has me worried.


Wasn't there a thread with a link that said bulldozer instructions were not built into Zen and will make it crash?

Re gcc-6.3.0:

I did it on (k10) by the old-fashioned, "you don't really have to do this" way.

Code:
emerge -1v gcc
emerge -1v libtool
emerge -1v glibc
emerge -e @system
emerge -e @world
I have --keep-going in my make,conf and had to do "emerge --resume" several times, but eventually they all compiled, which tells me that something is wrong with dependencies.

On one machine I used a script from khayyam that emerge -e world --excluding things in system, but I don't have it anymore. It's somewhere in this forum.
Straight out "emerge -e @world" took 11 or 12 hours on a Phenom II X6 8G memory, 10,000 RPM drive.
[/quote]
There was, thats why a few of us have tried to keep the gentoo wiki upto date.
Originally Haswell was recomended for gcc-5.3 but there were some segfaults.
CFLAGS="-O2 -march=bdver4 -mno-fma4 -mno-tbm -mno-xop -mno-lwp"

https://wiki.gentoo.org/wiki/Ryzen
_________________
The best argument against democracy is a five-minute conversation with the average voter
Great Britain is a republic, with a hereditary president, while the United States is a monarchy with an elective king
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3367
Location: Gainesville, Florida

PostPosted: Thu May 25, 2017 4:59 pm    Post subject: Reply with quote

Tony0945,
From https://wiki.gentoo.org/wiki/Ryzen#Kernel
Quote:
GCC 6.3+

GCC 6.3+ has support for the znver1 compiler optimization. For optimal performance, this can be enabled in make.conf.
FILE /etc/portage/make.confZen compiler optimization

CFLAGS="-O2 -march=znver1"

GCC 6.3 is not presently optimized for Ryzen. [1]
GCC 5.4

While GCC 5.4 does not support zen core specific optimization, -march=bdver4 has been shown to be functional and stable. However, since Zen dropped the instruction set extensions FMA4, TBM, XOP and LWP, they should be disabled accordingly:
FILE /etc/portage/make.confZen compiler optimization for GCC 5.4 and lower

CFLAGS="-O2 -march=bdver4 -mno-fma4 -mno-tbm -mno-xop -mno-lwp"

Note
Previously -march=haswell was said to be functional with Zen[2], but a Gentoo developer experienced various SEGVs with this option.
Note
The use of bare -march=bdver4 was said to be functional without issues, nevertheless it may still produce faulty code due to the lack of before mentioned instruction set extensions. Bulldozer has them, Zen does not.

Optional, but may produce better code: Add new instruction set extensions introduced with Zen individually (ADCX, RDSEED, MWAITX, SHA, CLZERO, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT), using -march=bdver4 (Bulldozer Version 4 i.e. Excavator) as the starting point:
FILE /etc/portage/make.confEXPERIMENTAL compiler optimization for GCC 5.4 specifying new extensions for Zen

CFLAGS="-O2 -march=bdver4 -mno-fma4 -mno-tbm -mno-xop -mno-lwp -mclzero -madx -mrdseed -mmwaitx -msha -mxsavec -mxsaves -mclfushopt -mpopcnt"

On my new non-x (so far) sda6 Gentoo install with kernel-4.11.1 and 2, and 12-rc2, gcc-6.3 ~amd64, it's been some sort of struggle with countless failed --keep-going & --resume --skipfirst, plus many, many /etc/portage/ file edits and then trial and error rebuild attempts. Mostly python, ruby, and openssl mysteries and conflict which for some reason won't resolve with the known fixes for such stuff. Makes me think the toolchain is suspect. There are also random failures on other system file rebuild packages.

I can confirm that with gcc-6.3.0 the -mclfushopt flag mentioned on the ryzen-wiki page completely breaks the gcc-6.3.0 compiler. Can't emerge anything- all packages immediately give this error:
Code:
gcc: error: unrecognized command line option ‘-mclfushopt’; did you mean ‘-mclflushopt’?
make: *** [<builtin>: glibc-test] Error 1
emake failed
 * Simple build failed ... assuming this is desired #324685
configure: error: C compiler cannot create executables
Remove the flag, and all is normal.

Just to clarify, my problems are on a fresh install, which I did chrooted into sda6 from my Mint-kde sda1 install.. Haven't gotten to trying to rebuild my sda11 install. Since you've done something similar, what's your opinion about chrooting into sda11 from mint, and doing the rebuild like that, as I mentioned above? I might avoid having to do another a fresh install (which hasn't gone well on sda6), or avoid moving the hard drive to the old AM3+ system for the rebuild, and back to the AM4 box.

Seems like chrooting to sda11, changing the flags to bdver4 + the extra zen flags mentioned on the wiki, and then rebuild @world might accomplish
the zen update. Am I missing a reason that won't work?
_________________
Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.32, gcc-10.2.0, kernel-5.9.1-gentoo, 5.10.0-rc1 USE=experimental
Back to top
View user's profile Send private message
ct85711
Veteran
Veteran


Joined: 27 Sep 2005
Posts: 1791

PostPosted: Thu May 25, 2017 6:41 pm    Post subject: Reply with quote

Quote:
I can confirm that with gcc-6.3.0 the -mclfushopt flag mentioned on the ryzen-wiki page completely breaks the gcc-6.3.0 compiler. Can't emerge anything- all packages immediately give this error:
Code:

gcc: error: unrecognized command line option ‘-mclfushopt’; did you mean ‘-mclflushopt’?
make: *** [<builtin>: glibc-test] Error 1
emake failed
 * Simple build failed ... assuming this is desired #324685
configure: error: C compiler cannot create executables   

Remove the flag, and all is normal.


I think this may be more of a case of a typo on the -mclfushopt...

Quote:
Optional, but may produce better code: Add new instruction set extensions introduced with Zen individually (ADCX, RDSEED, MWAITX, SHA, CLZERO, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT), using -march=bdver4 (Bulldozer Version 4 i.e. Excavator) as the starting point:
FILE /etc/portage/make.confEXPERIMENTAL compiler optimization for GCC 5.4 specifying new extensions for Zen

CFLAGS="-O2 -march=bdver4 -mno-fma4 -mno-tbm -mno-xop -mno-lwp -mclzero -madx -mrdseed -mmwaitx -msha -mxsavec -mxsaves -mclfushopt -mpopcnt"

Earlier you quoted the extension should be CLFLUSHOPT, but the CFLAG does not match the same name (taking off the -m) you put down clfushopt not clflushopt

Edit: Looking through some more, you may need to do the same some the ADCX extension too... In that you have -madx, when you may be meaning adcx instead?
Back to top
View user's profile Send private message
wrc1944
Advocate
Advocate


Joined: 15 Aug 2002
Posts: 3367
Location: Gainesville, Florida

PostPosted: Thu May 25, 2017 7:55 pm    Post subject: Reply with quote

ct85711,
Nope, can't be a typo- I copied the flags of the ryzen wiki page, and pasted them into my make.conf.

The gcc error I posted was a copy/paste out of the konsole terminal I was using in the Mint sda6 chroot.
I thought they made a typo mistake on the wiki page, so I just checked and it still is -mclfushopt.

Here's the current wiki segment:
Quote:
Optional, but may produce better code: Add new instruction set extensions introduced with Zen individually (ADCX, RDSEED, MWAITX, SHA, CLZERO, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT), using -march=bdver4 (Bulldozer Version 4 i.e. Excavator) as the starting point:
FILE /etc/portage/make.confEXPERIMENTAL compiler optimization for GCC 5.4 specifying new extensions for Zen

CFLAGS="-O2 -march=bdver4 -mno-fma4 -mno-tbm -mno-xop -mno-lwp -mclzero -madx -mrdseed -mmwaitx -msha -mxsavec -mxsaves -mclfushopt -mpopcnt"
Hmmmmm... So apparently the "typo" is actually on the wiki page itself- didn't notice before when I c/p'd.

You are right. They did omit the l after the f. Easy to make that type of typo- I do it myself all the time. Nice catch! :D

Wow. One missed character utterly destroys gcc! 8O
_________________
Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
Gentoo ~amd64 plasma, glibc-2.32, gcc-10.2.0, kernel-5.9.1-gentoo, 5.10.0-rc1 USE=experimental
Back to top
View user's profile Send private message
ct85711
Veteran
Veteran


Joined: 27 Sep 2005
Posts: 1791

PostPosted: Thu May 25, 2017 8:46 pm    Post subject: Reply with quote

Well, now we know the wiki page was wrong, and needed to be fixed :)

Though doing some research, it seems the mistake isn't completely due to the person that made the wiki page. According to gcc online manual for 6.3 and 7.1, they list the CPU option as -mclfushopt, which as you stated is what wiki says/said. So I think this is more of a an issue in gcc's official documentation than on our wiki...

https://gcc.gnu.org/onlinedocs/gcc-6.3.0/gcc/x86-Options.html#x86-Options
https://gcc.gnu.org/onlinedocs/gcc-7.1.0/gcc/x86-Options.html#x86-Options

As far as the -madx flag, I am not seeing this specified anywhere in the online documentation, though the docs list ADCX available on 5 different marches.... So I am thinking the -madx may be something that doesn't exist as a flag to enable? This may be something that we may need to go through gcc's source code and verify what is correct.
Back to top
View user's profile Send private message
Naib
Watchman
Watchman


Joined: 21 May 2004
Posts: 5855
Location: Removed by Neddy

PostPosted: Thu May 25, 2017 9:38 pm    Post subject: Reply with quote

is it wrong?

That option is under gcc-5.4 & is marked as experimental.
If you show the flags that gcc-6.3 or gcc-7.1 will use:

Quote:
echo | gcc -### -E - -march=native
Using built-in specs.
COLLECT_GCC=/usr/x86_64-pc-linux-gnu/gcc-bin/7.1.0/gcc
Target: x86_64-pc-linux-gnu
Configured with: /var/tmp/portage/sys-devel/gcc-7.1.0-r1/work/gcc-7.1.0/configure --host=x86_64-pc-linux-gnu --build=x86_64-pc-linux-gnu --prefix=/usr --bindir=/usr/x86_64-pc-linux-gngcc-bin/7.1.0 --includedir=/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/include --datadir=/usr/share/gcc-data/x86_64-pc-linux-gnu/7.1.0 --mandir=/usr/share/gcc-data/x86_64-pc-linux-gnu/7.1.man --infodir=/usr/share/gcc-data/x86_64-pc-linux-gnu/7.1.0/info --with-gxx-include-dir=/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/include/g++-v7 --with-python-dir=/share/gcc-data/x86_64--linux-gnu/7.1.0/python --enable-languages=c,c++,fortran --enable-obsolete --enable-secureplt --disable-werror --with-system-zlib --enable-nls --without-included-gettext --enable-checng=release --with-bugurl=https://bugs.gentoo.org/ --with-pkgversion='Gentoo 7.1.0-r1 p1.1' --disable-esp --enable-libstdcxx-time --enable-shared --enable-threads=posix --enable-__cxa_exit --enable-clocale=gnu --enable-multilib --with-multilib-list=m32,m64 --disable-altivec --disable-fixed-point --enable-targets=all --disable-libgcj --enable-libgomp --disable-libmulap --disable-libssp --disable-libcilkrts --disable-libmpx --enable-vtable-verify --enable-libvtv --enable-lto --without-isl --enable-libsanitizer --disable-default-pie --enable-defau-ssp
Thread model: posix
gcc version 7.1.0 (Gentoo 7.1.0-r1 p1.1)
COLLECT_GCC_OPTIONS='-E' '-march=native'
/usr/libexec/gcc/x86_64-pc-linux-gnu/7.1.0/cc1 -E -quiet - "-march=znver1" -mmmx -mno-3dnow -msse -msse2 -msse3 -mssse3 -msse4a -mcx16 -msahf -mmovbe -maes -msha -mpclmul -mpopcnt -mm -mno-lwp -mfma -mno-fma4 -mno-xop -mbmi -mno-sgx -mbmi2 -mno-tbm -mavx -mavx2 -msse4.2 -msse4.1 -mlzcnt -mno-rtm -mno-hle -mrdrnd -mf16c -mfsgsbase -mrdseed -mprfchw -madx -mfxsr -mave -mxsaveopt -mno-avx512f -mno-avx512er -mno-avx512cd -mno-avx512pf -mno-prefetchwt1 -mclflushopt -mxsavec -mxsaves -mno-avx512dq -mno-avx512bw -mno-avx512vl -mno-avx512ifma -mno-av12vbmi -mno-avx5124fmaps -mno-avx5124vnniw -mno-clwb -mmwaitx -mclzero -mno-pku -mno-rdpid --param "l1-cache-size=32" --param "l1-cache-line-size=64" --param "l2-cache-size=512" "-mtu=znver1"
COMPILER_PATH=/usr/libexec/gcc/x86_64-pc-linux-gnu/7.1.0/:/usr/libexec/gcc/x86_64-pc-linux-gnu/7.1.0/:/usr/libexec/gcc/x86_64-pc-linux-gnu/:/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/:/uslib/gcc/x86_64-pc-linux-gnu/:/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/../../../../x86_64-pc-linux-gnu/bin/
LIBRARY_PATH=/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/:/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/../../../../lib64/:/lib/../lib64/:/usr/lib/../lib64/:/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0./../../../x86_64-pc-linux-gnu/lib/:/usr/lib/gcc/x86_64-pc-linux-gnu/7.1.0/../../../:/lib/:/usr/lib/
COLLECT_GCC_OPTIONS='-E' '-march=native'


All someone has done is "backported" flags to compensate for the fact gcc-5.x does not have a znver1 march to auto-set them.
could gcc be wrong? sure
_________________
The best argument against democracy is a five-minute conversation with the average voter
Great Britain is a republic, with a hereditary president, while the United States is a monarchy with an elective king
Back to top
View user's profile Send private message
thumper
Guru
Guru


Joined: 06 Dec 2002
Posts: 537
Location: Venice FL

PostPosted: Thu May 25, 2017 9:44 pm    Post subject: Reply with quote

For one reason or another I don't want to build from scratch a new Gentoo installation having just copied my system to a new drive and moved on many times in the past, I'd prefer to continue the tradition.

So my question is, has anyone here tried recompiling some or all of their system to an apparent common denominator like ( -mtune=generic -march=x86-64 ) and then copying the system over to a Ryzen machine, and then recompile using the appropriate CFLAGS for Ryzen?

Thanks;

George
Back to top
View user's profile Send private message
likewhoa
l33t
l33t


Joined: 04 Oct 2006
Posts: 778
Location: Brooklyn, New York

PostPosted: Sun May 28, 2017 10:46 pm    Post subject: Reply with quote

thumper wrote:
For one reason or another I don't want to build from scratch a new Gentoo installation having just copied my system to a new drive and moved on many times in the past, I'd prefer to continue the tradition.

So my question is, has anyone here tried recompiling some or all of their system to an apparent common denominator like ( -mtune=generic -march=x86-64 ) and then copying the system over to a Ryzen machine, and then recompile using the appropriate CFLAGS for Ryzen?

Thanks;

George


that should work
Back to top
View user's profile Send private message
thumper
Guru
Guru


Joined: 06 Dec 2002
Posts: 537
Location: Venice FL

PostPosted: Mon May 29, 2017 12:20 am    Post subject: Reply with quote

likewhoa wrote:
thumper wrote:
For one reason or another I don't want to build from scratch a new Gentoo installation having just copied my system to a new drive and moved on many times in the past, I'd prefer to continue the tradition.

So my question is, has anyone here tried recompiling some or all of their system to an apparent common denominator like ( -mtune=generic -march=x86-64 ) and then copying the system over to a Ryzen machine, and then recompile using the appropriate CFLAGS for Ryzen?

Thanks;

George


that should work


I checked the cflags used to compile openssl on Centos and Ubuntu using the speed command, and that was enough to convince me it should work, I changed my make.conf accordingly and am about 90% done with the complete system emerge, most of my parts arrived today, memory and a few new case fans will be here Monday so I'll soon know for sure.. :)

Thanks;

George
Back to top
View user's profile Send private message
likewhoa
l33t
l33t


Joined: 04 Oct 2006
Posts: 778
Location: Brooklyn, New York

PostPosted: Mon May 29, 2017 4:36 am    Post subject: Reply with quote

thumper wrote:
likewhoa wrote:
thumper wrote:
For one reason or another I don't want to build from scratch a new Gentoo installation having just copied my system to a new drive and moved on many times in the past, I'd prefer to continue the tradition.

So my question is, has anyone here tried recompiling some or all of their system to an apparent common denominator like ( -mtune=generic -march=x86-64 ) and then copying the system over to a Ryzen machine, and then recompile using the appropriate CFLAGS for Ryzen?

Thanks;

George


that should work


I checked the cflags used to compile openssl on Centos and Ubuntu using the speed command, and that was enough to convince me it should work, I changed my make.conf accordingly and am about 90% done with the complete system emerge, most of my parts arrived today, memory and a few new case fans will be here Monday so I'll soon know for sure.. :)

Thanks;

George


With Ryzen I would recommend having the following:

  • Kernel >=4.8 (4.11 best)
  • >=GCC-6 (GCC-7 best)
  • -march=haswell for <=gcc-5
  • -march=znver1 for >=gcc-6 although with gcc-7 -march=native should work haven't tried
  • Make sure you RAM/CPU settings in BIOS are properly set, pay attention to memory timings and voltages
  • MAKEOPTS=-j16 of course :D


That should be it. My current setup is as follow

Quote:

dmesg | grep -i carbon; echo; uname -a; echo; /lib/libc.so.6
[ 0.000000] DMI: Micro-Star International Co., Ltd MS-7A32/X370 GAMING PRO CARBON (MS-7A32), BIOS 1.50 04/27/2017
[ 8.688089] Hardware name: Micro-Star International Co., Ltd MS-7A32/X370 GAMING PRO CARBON (MS-7A32), BIOS 1.50 04/27/2017

Linux lisa 4.8.17-hardened-r2 #8 SMP PREEMPT Fri May 26 21:35:45 EDT 2017 x86_64 AMD Ryzen 7 1700 Eight-Core Processor AuthenticAMD GNU/Linux

GNU C Library (Gentoo 2.23-r3 p7) stable release version 2.23, by Roland McGrath et al.
Copyright (C) 2016 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 7.1.0.
Available extensions:
C stubs add-on version 2.1.2
crypt add-on version 2.1 by Michael Glad and others
GNU Libidn by Simon Josefsson
Native POSIX Threads Library by Ulrich Drepper et al
BIND-8.2.3-T5B
libc ABIs: UNIQUE IFUNC
For bug reporting instructions, please see:
<https://bugs.gentoo.org/>.


F.Y.I I am currently stable at 3.9GHz 1.3626vCore 1.0750v NB and 3.5v DRAM at 3200MHz 14-14-14-30
Back to top
View user's profile Send private message
drizzt
Guru
Guru


Joined: 21 Jul 2002
Posts: 424

PostPosted: Mon May 29, 2017 4:34 pm    Post subject: Reply with quote

likewhoa wrote:
thumper wrote:
likewhoa wrote:
thumper wrote:
For one reason or another I don't want to build from scratch a new Gentoo installation having just copied my system to a new drive and moved on many times in the past, I'd prefer to continue the tradition.

So my question is, has anyone here tried recompiling some or all of their system to an apparent common denominator like ( -mtune=generic -march=x86-64 ) and then copying the system over to a Ryzen machine, and then recompile using the appropriate CFLAGS for Ryzen?

Thanks;

George


that should work


I checked the cflags used to compile openssl on Centos and Ubuntu using the speed command, and that was enough to convince me it should work, I changed my make.conf accordingly and am about 90% done with the complete system emerge, most of my parts arrived today, memory and a few new case fans will be here Monday so I'll soon know for sure.. :)

Thanks;

George


With Ryzen I would recommend having the following:

  • Kernel >=4.8 (4.11 best)
  • >=GCC-6 (GCC-7 best)
  • -march=haswell for <=gcc-5
  • -march=znver1 for >=gcc-6 although with gcc-7 -march=native should work haven't tried
  • Make sure you RAM/CPU settings in BIOS are properly set, pay attention to memory timings and voltages
  • MAKEOPTS=-j16 of course :D


That should be it. My current setup is as follow

Quote:

dmesg | grep -i carbon; echo; uname -a; echo; /lib/libc.so.6
[ 0.000000] DMI: Micro-Star International Co., Ltd MS-7A32/X370 GAMING PRO CARBON (MS-7A32), BIOS 1.50 04/27/2017
[ 8.688089] Hardware name: Micro-Star International Co., Ltd MS-7A32/X370 GAMING PRO CARBON (MS-7A32), BIOS 1.50 04/27/2017

Linux lisa 4.8.17-hardened-r2 #8 SMP PREEMPT Fri May 26 21:35:45 EDT 2017 x86_64 AMD Ryzen 7 1700 Eight-Core Processor AuthenticAMD GNU/Linux

GNU C Library (Gentoo 2.23-r3 p7) stable release version 2.23, by Roland McGrath et al.
Copyright (C) 2016 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 7.1.0.
Available extensions:
C stubs add-on version 2.1.2
crypt add-on version 2.1 by Michael Glad and others
GNU Libidn by Simon Josefsson
Native POSIX Threads Library by Ulrich Drepper et al
BIND-8.2.3-T5B
libc ABIs: UNIQUE IFUNC
For bug reporting instructions, please see:
<https://bugs.gentoo.org/>.


F.Y.I I am currently stable at 3.9GHz 1.3626vCore 1.0750v NB and 3.5v DRAM at 3200MHz 14-14-14-30


march=haswell seems not to be the best option see here and https://wiki.gentoo.org/wiki/Ryzen
_________________
People don't have to earn my respect. I offer my respect to them, but be careful to lose my respect...
Back to top
View user's profile Send private message
likewhoa
l33t
l33t


Joined: 04 Oct 2006
Posts: 778
Location: Brooklyn, New York

PostPosted: Mon May 29, 2017 5:20 pm    Post subject: Reply with quote

drizzt wrote:

march=haswell seems not to be the best option see here and https://wiki.gentoo.org/wiki/Ryzen


Worked for me on gcc-5 but i didn't have many packages to begin with but some major ones like thunderbird, vlc, libreoffice, chromium, fluxbox, xfce4 all compiled fine with it. My problems started when I moved to gcc-6
Back to top
View user's profile Send private message
Chewi
Developer
Developer


Joined: 01 Sep 2003
Posts: 881
Location: Edinburgh, Scotland

PostPosted: Fri Jun 02, 2017 8:35 pm    Post subject: Reply with quote

I've recently joined this club. :D I have been having problems though. My previous CPU was a Nehalem i7 and I had been using -march=nehalem. I haven't rebuilt much of the system at all yet but userspace has been fine. It's the kernel that has been randomly freezing on me when idle. I built 4.11.3 with GCC 6.3.0 and -march=native. Since selecting "Generic-x86-64" instead, it's been up for 13 hours and counting. Opinions on whether native works with 6.3.0 appear mixed but I would say not. I guess the same goes for znver1. I'll give bdver4 a try. The Gentoo kernel patch that adds bdver4 doesn't add the other required flags but you can pass them using KCFLAGS.
Back to top
View user's profile Send private message
Naib
Watchman
Watchman


Joined: 21 May 2004
Posts: 5855
Location: Removed by Neddy

PostPosted: Fri Jun 02, 2017 8:45 pm    Post subject: Reply with quote

Chewi wrote:
I've recently joined this club. :D I have been having problems though. My previous CPU was a Nehalem i7 and I had been using -march=nehalem. I haven't rebuilt much of the system at all yet but userspace has been fine. It's the kernel that has been randomly freezing on me when idle. I built 4.11.3 with GCC 6.3.0 and -march=native. Since selecting "Generic-x86-64" instead, it's been up for 13 hours and counting. Opinions on whether native works with 6.3.0 appear mixed but I would say not. I guess the same goes for znver1. I'll give bdver4 a try. The Gentoo kernel patch that adds bdver4 doesn't add the other required flags but you can pass them using KCFLAGS.
Care to add yourself to this survey: Gentoo & Ryzen config & stability questionair (public) also: the auto-updating sheet of the results
Just trying to help pull together the pieces of info to help
_________________
The best argument against democracy is a five-minute conversation with the average voter
Great Britain is a republic, with a hereditary president, while the United States is a monarchy with an elective king
Back to top
View user's profile Send private message
Display posts from previous:   
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3 ... 15, 16, 17 ... 20, 21, 22  Next
Page 16 of 22

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum