ChanServ changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630)
<steev>
bamse: so, i was compiling a rather large rust app... and had something happen - not sure what exactly... but ufs was again being very unhelpful (and mounted RO) so no log... however... upon reboot, even though i'm on 5.12.19... i'm seeing that same rcg clock complaint that I was seeing in 5.13
<derzahl>
wow, 5.12.19 took my compile time of 5.12.19 down from 34m35s to 26m9s
<derzahl>
you werent joking about a 30%+ improvement
derzahl has quit [Ping timeout: 480 seconds]
FizzBuzz has quit []
derzahl has joined #aarch64-laptops
pinskia has quit [Remote host closed the connection]
pinskia has joined #aarch64-laptops
iivanov has joined #aarch64-laptops
derzahl has quit [Ping timeout: 480 seconds]
derzahl has joined #aarch64-laptops
FizzBuzz has joined #aarch64-laptops
<mattst88>
derzahl: the 30% improvement was from using ext4 rather than zfs?
<robclark>
mattst88: there should be some emmc throughput improvements in the pipe.. maybe v5.15?.. (which should help lazor.. c630 is ufs, but I suppose might also benefit)
<mattst88>
ah, nice
<robclark>
I think it was mostly iommu non-strict..
<robclark>
actually, I think you can get that already w/ addition of kernel cmdline thing.. I think iommu.strict=0 or something along those lines
<robclark>
(iirc, the thing that is in the works was a thing to set that more precisely per device, rather than globally)
<steev>
mattst88: the 30% is because i added the lmh driver to my 5.12
<steev>
i still need to play with the component patches and the edp-panel patches, but they need a bit of attention and i just didn't have it last night at 2am
<mattst88>
what is lmh?
<robclark>
oh, 845/850 was being artificially limited?
<robclark>
lmh is apparently "Limits Management Hardware"
<mattst88>
oh, I imagine this is similar to Intel's DPTF or something?
<robclark>
yeah, not sure how DPTF works (ie. how much is hw vs fw).. but I guess at least same idea.. https://lwn.net/Articles/860817/
<robclark>
I don't *think* it should matter too much for 7c.. there is some last-resort thermal limit stuff configured by fw before kernel starts but I've never seen it triggered (or even really, and thermal throttling triggered)
<steev>
si, it's still a wee bit buggy (crashes if the cpus are busy and the system suspends/sleeps) but overall, the performance bump is very nice
iivanov has quit [Remote host closed the connection]