On 2020-02-29 16:01, Marcus Meissner wrote:

On Sat, Feb 29, 2020 at 02:00:37PM +0100, Frans de Boer wrote:

...

LS,

Just installed the beta version of leap15.2, just to see if there is anything too upgrade for. Nope, just a new kernel and may some other new or updated packages, but even glibc stays ancient, just as many other packages.  Thus keeping leap15.2 as slow as leap15.1 which is 4 times slower then TW for mathematics due to the ancient glibc library and possible the very ancient gcc tool set.

But, maybe I am pessimistic and can someone correct me? It sounds like you would be more happy with Tumbleweeds bleeding edge?

Ciao, Marcus

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages. The thing is, I do like to have a stable distribution in case TW sometimes misfired. But using a distro which uses 3-4- years old packages - and after 8 months being even 4-5 years or more - can only be used as a last resort. --- Frans. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 29.02.20 um 16:22 schrieb Frans de Boer:

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages.

The thing is that much effort is saved by basing (in this case) 15.2 just on SLES15-SP2, and letting SUSE do the maintenance on many of the base packages instead of assigning the community the task of maintaining it.

I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages.

Nobody is reverting to anything, it is just the saved effort of updating *and keeping the updated packages secure*. We had this before Leap, until 13.2. The thing was exactly as you described it: fork off from Factory/Tumbleweed at some point in time and then maintain it for $SOME years. The problem with that is lack of manpower and expertise.

The thing is, I do like to have a stable distribution in case TW sometimes misfired. But using a distro which uses 3-4- years old packages - and after 8 months being even 4-5 years or more - can only be used as a last resort.

I have not really a problem running my stuff on Leap. And I'm pretty happy that I (as a packager / community member) had almost zero work to do for 15.2, because I did not need to touch the packages. And I almost never have to care for "old crap", I usually only update tumbleweed to the latest and greatest. I very much like it like that. Did you really measure the 4x slowdown of Leap vs TW or is this just a guess based on the software version? Often the SLES versions get fixes / improvements backported from newer releases, but keep the compatibility (and version numbers) with the old version. If you notice such big performance differences in Leap vs tumbleweed, then filing a bug might be appropriate. SUSE might be interested in porting back these performance fixes to SLES15-SP2 (from there Leap 15.2 will get these more or less automatically). Have fun, -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 29.02.20 um 17:24 schrieb Jan Engelhardt:

On Saturday 2020-02-29 16:22, Frans de Boer wrote:

...

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages.

You talk the talk, but when was the last time you *actually* needed a glibc package right out of the factory?

I tried to install the new SAGE math package some weeks ago, which didn't work since glibc version was too ancient. But for other things LEAP works just fine, so I don't need to mess around with TW. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 29.02.20 um 18:23 schrieb Stefan Seyfried:

Am 29.02.20 um 17:49 schrieb zb4ng:

...

I tried to install the new SAGE math package some weeks ago, which didn't work since glibc version was too ancient. But for other things LEAP works just fine, so I don't need to mess around with TW.

More details, please, so that we can reproduce / advise possible workarounds.

For information, SAGE (https://www.sagemath.org/) is a huge python based software; they even include their own python interpreter (which is still python 2.x last time I checked). I tried the binary package (Sage 8.9) and also to compile the source package, both without success. After some fiddling, I decided to stick with the older version I still had and deleted the new one, so, unfortunately, I can't give exact details. Maybe it can be fixed it with some environment variables. I'm going to wait until LEAP 15.2 comes out and then I'll try it again and give you details if it still doesn't work. Though maybe I better use their docker files, if I can figure out how that works! -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Saturday 2020-02-29 21:14, Stefan Seyfried wrote:

...

For information, SAGE (https://www.sagemath.org/) is a huge python based software; they even include their own python interpreter (which is still python 2.x last time I checked).

I just tried it.

I downloaded the binary package sage-9.0-Debian_GNU_Linux_10-x86_64.tar.bz2, extracted it, run "./sage" and it failed for too old glibc.

Builds produced under a new toolchain do not necessarily work in an old environment. This is has been known for decades, which is why software where the vendor tries to make some kind of "universal" binary, simply builds it using a sufficiently _old_ toolchain, e.g. something like RHEL6. You would not expect Win3.1 to run just any Win95 program either. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 29.02.20 um 21:14 schrieb Stefan Seyfried:

Am 29.02.20 um 20:50 schrieb zb4ng:

...

Am 29.02.20 um 18:23 schrieb Stefan Seyfried:

...

Am 29.02.20 um 17:49 schrieb zb4ng:

...

I tried to install the new SAGE math package some weeks ago, which didn't work since glibc version was too ancient. But for other things LEAP works just fine, so I don't need to mess around with TW.

More details, please, so that we can reproduce / advise possible workarounds.

For information, SAGE (https://www.sagemath.org/) is a huge python based software; they even include their own python interpreter (which is still python 2.x last time I checked).

I just tried it.

I downloaded the binary package sage-9.0-Debian_GNU_Linux_10-x86_64.tar.bz2, extracted it, run "./sage" and it failed for too old glibc. But debian 10 is brand new, so next try: download sage-9.0-Debian_GNU_Linux_9-x86_64.tar.bz2, extract it, run "./sage" and it seems to work fine.

It works for me, too (including Jupyter)! Somehow, it never occurred to me before to try the Debian versions. Thanks! -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dne 29. 02. 20 v 17:24 Jan Engelhardt napsal(a):

On Saturday 2020-02-29 16:22, Frans de Boer wrote:

...

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages. You talk the talk, but when was the last time you *actually* needed a glibc package right out of the factory? I currently use glibc from factory for one game: https://bugzilla.opensuse.org/show_bug.cgi?id=1163301

I expect that problem will widen. At least games already started to abandoning glibc older then Ubuntu's 18.04. Chrome uses usually not old glibc, so when Chrome updates glibc dependencies again, we will stuck not only without proprietary version of Chrome, but without all electron based applications. I don't know, why someone decided not to upgrade glibc. We have newer glibc in SLE 11 SP2 (or SP3?) and in 12 SP2/Leap 42.2, so I expected this update for Leap 15.2 . For desktop are newer drivers (backported or within newer kernel, mesa, alsa) and newer glibc needed. First thing Leap 15.2 has, second thing is missing. Even worse is, that old glibc in 15.2 means very probably old glibc in 15.3, because 15.2 is that "refresh" version, where bigger updates were expected. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dne 01. 03. 20 v 22:12 Michal Kubecek napsal(a):

On Sun, Mar 01, 2020 at 11:26:10AM +0100, Daniel Noga wrote:

...

Even worse is, that old glibc in 15.2 means very probably old glibc in 15.3, because 15.2 is that "refresh" version, where bigger updates were expected. Interesting. Where does this information come from?

Michal Kubecek

Leap is alligned to SLE and according to https://www.suse.com/betaprogram/sle-beta/ SLE 15 SP2 is: "SUSE Linux Enterprise 15 Service Pack 2 (SLE 15 SP2) is the a Refresh Service Pack." Description what is difference between "Consolidation service pack" and "Refresh service pack" was mentioned in some conference, I think. Unfortunatelly I cannot find it on YouTube. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Sunday 2020-03-01 11:26, Daniel Noga wrote:

...

...

in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages. You talk the talk, but when was the last time you *actually* needed a glibc package right out of the factory? I currently use glibc from factory for one game: https://bugzilla.opensuse.org/show_bug.cgi?id=1163301

So that is a packaging problem of the game then. They cannot expect everyone has a particular glibc version, so either they rebuild it for an older glibc, bundle libc (applicable to any other library), or just outright ship source code so it can be built for whatever libc the user is running. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 29/02/2020 16.22, Frans de Boer wrote:

On 2020-02-29 16:01, Marcus Meissner wrote:

...

On Sat, Feb 29, 2020 at 02:00:37PM +0100, Frans de Boer wrote:

...

LS,

Just installed the beta version of leap15.2, just to see if there is anything too upgrade for. Nope, just a new kernel and may some other new or updated packages, but even glibc stays ancient, just as many other packages. Thus keeping leap15.2 as slow as leap15.1 which is 4 times slower then TW for mathematics due to the ancient glibc library and possible the very ancient gcc tool set.

But, maybe I am pessimistic and can someone correct me? It sounds like you would be more happy with Tumbleweeds bleeding edge?

Ciao, Marcus

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages.

But the thing is, it is done basically by taking a snapshot of factory at some point and keeping it. And it is done by SLES, so Leap simply piggybacks. It is not done by reverting to 3-4 years old packages. - -- Cheers / Saludos, Carlos E. R. (from oS Leap 15.0 x86_64 (Minas Tirith)) -----BEGIN PGP SIGNATURE----- iHUEAREIAB0WIQQt/vKEw5659AgM/X2NrxRtxRYzXAUCXlqudAAKCRCNrxRtxRYz XK+CAP0UiZRdTakgCjwku5y18OhDtGlOCcSBYovqgAuV7YoF6wD7BJfy8I6X6zXa uiT6YxidsIPAWR5PmQkiFykggtPIdCc= =3z8K -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 29/02/2020 19.51, Hans-Peter Jansen wrote:

Am Samstag, 29. Februar 2020, 16:22:01 CET schrieb Frans de Boer:

...

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages.

The thing is, I do like to have a stable distribution in case TW sometimes misfired. But using a distro which uses 3-4- years old packages - and after 8 months being even 4-5 years or more - can only be used as a last resort.

I beg to differ, Frans, efforts in openSUSE are relatively good managed, with a low waste factor.

I'm a TW lover, and push it even to some of my customer's desktops/notebooks (those, who can deal with strong moving targets, and want to use Linux in the first place).

OTOH, Leap is *perfect* for servers, and that other part of users, that still wants to use a sane OS.

Nobody wants to deal with the upgrade fallout in complex setups (file, mail, dns, dhcp, ldap, smb, databases, vpn, telephony, obs, just to name a few) on a *regular* base, but use a stable *and* supported base system. And that's, where Leap excels. Even release upgrades are mostly smooth, and just a little sed'ing and one zyp away.

Leap offers stability and minimum fuzz with upgrades and the like, thanks to the shared base with SLE (keeping essential elements stable and supported), while TW offers the latest and greatest, while putting *a lot of* effort into testing. And thanks to OBS, we have a release collaboration tool with a pretty low friction level.

You can roll your own distribution, and (re-)package everything, that you don't like. glibc is a little exception of course, since adapting Leap to a new glibc, while possible, is not sensible. It will result in a new distribution, where you're on your own, and raging bitrot is no fun (been there, done that..).

If you need a new glibc, fine, use TW. And get everything else for free (current kernel, other libs, MESA, DEs just to name a few). Any hiccups are usually a matter of days to get fixed, but the base is stable and sound.

Hence, from my point of view, this distribution is targeted on real use cases, and does that very well. There are areas, that deserve improvements, but overall, the offerings from this project are on a very high quality level.. While the openSUSE project wasn't always positioned luckily in the past, it is in a very good shape as of now, with a firm standing, and sane prospects over all (and compared to others). I agree completely.

I am a desktop power user, and my needs are covered by Leap quite well. Sometimes I would like newer versions, but I don't like being on the edge. Thus my choice (given what is available) is Leap. Now, the claim that maths are four time slower I think needs study. - -- Cheers / Saludos, Carlos E. R. (from oS Leap 15.0 x86_64 (Minas Tirith)) -----BEGIN PGP SIGNATURE----- iHUEAREIAB0WIQQt/vKEw5659AgM/X2NrxRtxRYzXAUCXlq2gQAKCRCNrxRtxRYz XEsXAQCWnmZkWnEpQVfdPIpNUYhRfcNXIhKXwvs53MdqVSsUogD/RDZx89/s1/+I lkjwJbJOq3szMH2FKNVupYA0mPYUSfg= =xik+ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Frans de Boer [02-29-20 14:25]: [...]

I understand fully what you mean. Yes, it is a challenge to create a new distribution, it requires a lot of work. And creation is just one thing, next is the maintenance. The latter is the reason I do not build my own distro anymore.

All said and understood, does someone knows a stable distro which is somewhere between TW as bleeding edge and leap15.x? I did not followed the changing distro landscape lately.

but Tumbleweed is stable and tested and perfectly usable. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2020-02-29 20:59, Carlos E. R. wrote:

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256

On 29/02/2020 20.44, Patrick Shanahan wrote:

...

* Frans de Boer <> [02-29-20 14:25]: [...]

...

I understand fully what you mean. Yes, it is a challenge to create a new distribution, it requires a lot of work. And creation is just one thing, next is the maintenance. The latter is the reason I do not build my own distro anymore.

All said and understood, does someone knows a stable distro which is somewhere between TW as bleeding edge and leap15.x? I did not followed the changing distro landscape lately. but Tumbleweed is stable and tested and perfectly usable.

Sure? Didn't the config change off etc escape the testing? nfs? :-D

- -- Cheers / Saludos,

Carlos E. R.

(from oS Leap 15.0 x86_64 (Minas Tirith)) -----BEGIN PGP SIGNATURE-----

iHUEAREIAB0WIQQt/vKEw5659AgM/X2NrxRtxRYzXAUCXlrChQAKCRCNrxRtxRYz XGLyAP9Etbg1/wERdqstg+n1gf5TFTtsRCm8xaLnDnXvjg0q3wD+NZYU+cxmNKP8 tj3PeEyykuZe5lXzVj9KgW8mbd95+Ac= =zzuP -----END PGP SIGNATURE-----

Just a recent example by Carlos ;) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 29.02.20 20:59, Carlos E. R. wrote:

...

but Tumbleweed is stable and tested and perfectly usable.

Sure? Didn't the config change off etc escape the testing? nfs? :-D

it is!! i have used suse from 9.2 to 11.4evergreen and the updates have made "same" or even more problems (never have made only one system upgrade without loosing some main functions (xorgmonitorsetup, cups, nfs, samba....)) from one version to the next. at the moment i upgrade tumbleweed every about 2 month (with tumbleweed-cli) and sometimes i get a problem. but before, i got a lot of problems AT ONCE. ok - the last step at the moment "etc" i have not upgraded. it seems to me i have first to check 36 :-(( config.rpmnew files, - never cared before... :-) so for me tumbleweed is a wonderful work, 100 percent in work-environment usable. thank you all who made this possible!! simoN -- www.becherer.de -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Gesendet: Samstag, 29. Februar 2020 um 21:30 Uhr Von: "Simon Becherer" An: opensuse-factory@opensuse.org Betreff: Re: [opensuse-factory] why upgrading too leap15.2?

On 29.02.20 20:59, Carlos E. R. wrote:

...

...

but Tumbleweed is stable and tested and perfectly usable.

Sure? Didn't the config change off etc escape the testing? nfs? :-D

[...]

ok - the last step at the moment "etc" i have not upgraded. it seems to me i have first to check 36 :-(( config.rpmnew files, - never cared before... :-)

so for me tumbleweed is a wonderful work, 100 percent in work-environment usable. thank you all who made this possible!!

Fully support this view. I upgrade my TW installation only if there is interesting new stuff or new KDE versions, and I did not notice the /etc change at all (and still have some rpmnew files flying around, where I just learned now what they are good for) So, nothing is eaten as hot as it is cooked, and not every user is affected by every change. In general TW works perfect for me. Cheers Axel -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 29.02.20 um 20:24 schrieb Frans de Boer:

I understand fully what you mean. Yes, it is a challenge to create a new distribution, it requires a lot of work. And creation is just one thing, next is the maintenance. The latter is the reason I do not build my own distro anymore.

All said and understood, does someone knows a stable distro which is somewhere between TW as bleeding edge and leap15.x? I did not followed the changing distro landscape lately.

Why not use Leap as base and a TW container for applications where glibc performance matters? Greetings, Stephan -- Lighten up, just enjoy life, smile more, laugh more, and don't get so worked up about things. Kenneth Branagh -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 29. Februar 2020 16:22:01 MEZ schrieb Frans de Boer :

On 2020-02-29 16:01, Marcus Meissner wrote:

...

On Sat, Feb 29, 2020 at 02:00:37PM +0100, Frans de Boer wrote:

...

LS,

Just installed the beta version of leap15.2, just to see if there is anything too upgrade for. Nope, just a new kernel and may some other new or updated packages, but even glibc stays ancient, just as many other packages.  Thus keeping leap15.2 as slow as leap15.1 which is 4 times slower then TW for mathematics due to the ancient glibc library and possible the very ancient gcc tool set.

But, maybe I am pessimistic and can someone correct me? It sounds like you would be more happy with Tumbleweeds bleeding edge?

Ciao, Marcus

Very much more happy. I just don't understand why so much effort is made - or in my view energy wasted - to make a "new" distribution, based on years of old/ancient (core) packages. I, however, can understand that one would take a snapshot of TW and iron out some bugs and let that live for some time. That sounds more productive then reverting to 3-4 years old packages.

The thing is, I do like to have a stable distribution in case TW sometimes misfired. But using a distro which uses 3-4- years old packages - and after 8 months being even 4-5 years or more - can only be used as a last resort.

I agree and I'm not really happy with leap at moment. Many leap packages are much to old. I want a distribution between leap and Tumbleweed. A stable distri with some newer packages. Tumbleweed is to new and in my opinion not for production. Last example for this is the change with "etc". Leap 15.0, 15.1 and 15.2 are rather the same. All have the same old packages. Package which are divers years old. And leap is even to old for divers new server applications. I think the way, the strict "it is not in sle", is wrong. My opinion. Regards Eric -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2020年3月1日日曜日 17時43分54秒 CET AW wrote:

But I ran into this bug. To avoid such surprises, I wonder if I can install each single TW snapshots deliberately. I'd just wait one or two weeks if there is fallout.

It seems possible with TW-cli: https://nam05.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.co m%2Fboombatower%2Ftumbleweed-cli&data=02%7C01%7C%7C4d197e56288c468500010 8d7bdffe0d0%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C637186778985418970& amp;sdata=iFuJlbA%2Fb%2BA54o%2BwTWkewaqEZ24XnMCxEQ09ALHu4OA%3D&reserved= 0

Does anyone work with TW-cli?

I do. I find it perfect to avoid updating when broadcom-wl from Packman didn't get rebuilt for the new kernel for some reason while also being able to install new packages without worrying about possible binary incompatibilities. And the utility is in Tumbleweed itself, so you can use zypper to install it out-of-the-box. Regards Radosław Wyrzykowski -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hello, On Tue, Mar 03 2020, David C. Rankin wrote:

On 02/29/2020 07:00 AM, Frans de Boer wrote:

...

LS,

Just installed the beta version of leap15.2, just to see if there is anything too upgrade for. Nope, just a new kernel and may some other new or updated packages, but even glibc stays ancient, just as many other packages.  Thus keeping leap15.2 as slow as leap15.1 which is 4 times slower then TW for mathematics due to the ancient glibc library and possible the very ancient gcc tool set.

Feel free to install a newer gcc from https://build.opensuse.org/project/show/devel:gcc

...

But, maybe I am pessimistic and can someone correct me?

Chuckling.... can somebody correct you?

That's a very difficult question to answer. However, there are some observations to be made, "slower" and "faster" are qualitative and 4 times a qualitative measure is still itself qualitative.

Can you include some specific benchmarks you are referring to? The reason being, is I haven't seen any great leap in performance tied to glibc that I can think of.

From release notes of glibc 2.27:

* Optimized x86-64 asin, atan2, exp, expf, log, pow, atan, sin, cosf, sinf, sincosf and tan with FMA, contributed by Arjan van de Ven and H.J. Lu from Intel. * Optimized x86-64 trunc and truncf for processors with SSE4.1. * Optimized generic expf, exp2f, logf, log2f, powf, sinf, cosf and sincosf.

From release notes of glibc 2.29:

* Optimized generic exp, exp2, log, log2, pow, sinf, cosf, sincosf and tanf. About a year ago I was comparing 2.26 from SLES 15.0 and 2.29(*) on SPEC FPrate 2017 using gcc 9.1 (rc) on a Zen1-based AMD CPU. At -Ofast and with -march=native: - 521.wrf improved by 29% - 227.cam4 improved by 26% - 544.nab_r improved by 13.3% - 554.roms_r improved by almost 9% At -O2 (not a very good optimization level for SPEC rate) and generic march/mtune 503.bwases improved by 42%(!). It should not be difficult to construct a synthetic benchmark that would be 4x as fast with current glibc compared to 2.26 - just benchmark repeated invocations of exp :-) Some of these improvements - some of those coming from glibc 2.27 - were backported to SLE 15.1 and thus to leap 15.1. Unfortunately, the second batch was also quite significant and is not present in Leap glibc. (*) This 2.29 glibc also had a backport of /usr/include/finclude/math-vector-fortran.h which IIRC is upstream only since glibc 2.30. To make use of it, you need gcc 9.1 or newer.

Were it not for the cessation of updates, I could happily still be running 11.0 and the performance of assembly, C, C++ would still be very comparable. (though STL availability would be a problem)

I have repeatedly measured that a lot of CPU-intensive code produced with more modern toolchain has significantly better performance in the last few years, let alone since 11.0. Martin -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org