View previous topic :: View next topic |
Author |
Message |
mrbassie l33t
Joined: 31 May 2013 Posts: 772 Location: over here
|
Posted: Sat Aug 10, 2013 4:05 pm Post subject: Atom N270 |
|
|
lspci shows two cores although apparently this chip only has one with hyperthreading(?)
So am I correct that it should be -j2 and not -j3 under MAKEOPTS? Would it make any real difference either way?
Still a newb to this forum so sorry If I should have bumped an old thread instead of starting this one. |
|
Back to top |
|
|
eccerr0r Watchman
Joined: 01 Jul 2004 Posts: 9679 Location: almost Mile High in the USA
|
Posted: Sat Aug 10, 2013 4:46 pm Post subject: |
|
|
Try it both ways, depends on the workload which is better. I just use -j2 though, it also depends on how much RAM you have. I use -j8 on my i7 despite it only having 4 physical cores plus dual threading, some people say 9 is the best but it's very dependent on load, who is to say that 10 may be optimal to make sure there is no idle time? _________________ Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
Back to top |
|
|
mrbassie l33t
Joined: 31 May 2013 Posts: 772 Location: over here
|
Posted: Sat Aug 10, 2013 5:13 pm Post subject: |
|
|
only 2gigs ram.
guess 'll just change it and see if there's a noticable difference.
thanks for the input
EDIT: -j3 is the way to go, with -j2 compiling was a lot slower. |
|
Back to top |
|
|
marmoras n00b
Joined: 12 May 2012 Posts: 14
|
Posted: Mon Aug 26, 2013 9:04 pm Post subject: |
|
|
I use a -j3 on a compaq mini (buffed up with 2gb run), haven't noticed a problem. |
|
Back to top |
|
|
eccerr0r Watchman
Joined: 01 Jul 2004 Posts: 9679 Location: almost Mile High in the USA
|
Posted: Mon Aug 26, 2013 9:11 pm Post subject: |
|
|
It's amazing how much hyperthreading actually helps it seems?
How much of a difference did you see, what did you try building?
I don't have my eeePC 900A and honestly speaking I never tried -j3 as memory pressure (specifically onchip memory cache) was a big worry. Maybe it's not as much as one as I thought. _________________ Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
Back to top |
|
|
marmoras n00b
Joined: 12 May 2012 Posts: 14
|
Posted: Mon Aug 26, 2013 9:24 pm Post subject: |
|
|
Actually, I have compiled a full desktop (mostly out of curiosity); it took a couple of days. I use bins for libreoffice and firefox. After that it was not that bad with updating periodically, worst case scenario I leave it overnight. Compared to the main rig that has a 2600K, you ll get disheartened to compile anything on it. |
|
Back to top |
|
|
eccerr0r Watchman
Joined: 01 Jul 2004 Posts: 9679 Location: almost Mile High in the USA
|
Posted: Mon Aug 26, 2013 10:09 pm Post subject: |
|
|
Then the question also remains... what is your -j on your i7-2600K :) I suspect 4 is too few, and I don't think I'd try 12.
I didn't try playing with it on my current compute box - I just stuck it at 8 and left it... Not all ebuilds support parallel builds anyway I suppose. And yes the i7 is pretty much blowing away all my other machines, even my Core2 Quad, by a good margin. The SSD probably contributes however...
(My Asus eeePC Atom N270 has a non-default SSD in it, which helps quite a bit during compiling. Can't exactly help against true compute speed but unpacking into ramdisk and writing back to disk is much faster than with a mechanical disk (or even the disk that it came with. That POS 4GB SSD card writes so slow...)) _________________ Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
Back to top |
|
|
marmoras n00b
Joined: 12 May 2012 Posts: 14
|
Posted: Tue Aug 27, 2013 5:53 pm Post subject: |
|
|
On the i7 2600K (hyperthread) I used -j9, 2xphys cores +1, but I haven't given it much thought; nor will I, because it's very fast and it doesn't bother me. About the SSD, it might be a problem in the long term with the writes of the temp files. I haven't owned an ssd so really I don't know but may other people can give you insights. |
|
Back to top |
|
|
eccerr0r Watchman
Joined: 01 Jul 2004 Posts: 9679 Location: almost Mile High in the USA
|
Posted: Tue Aug 27, 2013 6:29 pm Post subject: |
|
|
From my machine (the i7-2700K) the 8GB of RAM I think is the biggest contributor since I build on tmpfs. But when it comes time to updating the disk (also when reading the tarball from disk, as well as dealing with the portage tree) the ssd makes a huge difference.
Despite running Gentoo, I haven't eaten that much life out of either of the two SSDs that I can monitor erase cycles (the 180G in my i7 and a 128G in my i5), I hope the 32G in my eeePC is similar. I expect the 32G to fail first though, just because it's older.
Still somewhat interesting to squeeze every bit out of it that I can. Despite the Atom being quite slow, the eeePC with its SSD keeps up with my p3-class boxes that have slow mechanical hard drives in them... _________________ Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
Back to top |
|
|
mrbassie l33t
Joined: 31 May 2013 Posts: 772 Location: over here
|
Posted: Tue Aug 27, 2013 7:03 pm Post subject: |
|
|
eccerr0r wrote: | It's amazing how much hyperthreading actually helps it seems?
How much of a difference did you see, what did you try building?
I don't have my eeePC 900A and honestly speaking I never tried -j3 as memory pressure (specifically onchip memory cache) was a big worry. Maybe it's not as much as one as I thought. |
Me? It was firefox, and the difference was 8-9 hours (ish) compared to 35-40 hours. |
|
Back to top |
|
|
eccerr0r Watchman
Joined: 01 Jul 2004 Posts: 9679 Location: almost Mile High in the USA
|
Posted: Tue Aug 27, 2013 7:17 pm Post subject: |
|
|
Hmm. I don't think the -j2 to -j3 change could have caused that much of a jump, something else must have changed... I could expect it to drop if the machine needed to swap with -j3, but an increase in speed? Something strange going on here. _________________ Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
Back to top |
|
|
mrbassie l33t
Joined: 31 May 2013 Posts: 772 Location: over here
|
Posted: Tue Aug 27, 2013 7:26 pm Post subject: |
|
|
That is genuinely the only change I made, it was an update I should mention. |
|
Back to top |
|
|
depontius Advocate
Joined: 05 May 2004 Posts: 3509
|
Posted: Wed Aug 28, 2013 11:43 am Post subject: |
|
|
eccerr0r wrote: | Despite running Gentoo, I haven't eaten that much life out of either of the two SSDs that I can monitor erase cycles (the 180G in my i7 and a 128G in my i5), I hope the 32G in my eeePC is similar. I expect the 32G to fail first though, just because it's older. |
How do you measure SSD life? I presume it's with "smartctl -a /dev/sda", but which field are you looking at? _________________ .sigs waste space and bandwidth |
|
Back to top |
|
|
eccerr0r Watchman
Joined: 01 Jul 2004 Posts: 9679 Location: almost Mile High in the USA
|
Posted: Wed Aug 28, 2013 12:11 pm Post subject: |
|
|
depontius wrote: | How do you measure SSD life? I presume it's with "smartctl -a /dev/sda", but which field are you looking at? |
It depends on the disk's firmware/brand, and the type of available data can differ.
If you have a Crucial M4 or Intel SSD 520 I found the fields needed, they both erasure count and "life remaining" which is extrapolated by the number of times they've been erased. Apparently in the newest smartmontools this should be listed in field 202 for the M4, and 233 for the Intel. Look at the "Value" field and compare it to the "Threshold" to see how much life you have left.
The M4 dumps its average erase count on field 173. My M4 has its raw number at 4 currently, which means it's been completely erased 4 times for the amount of time I've had this disk, which is about a year now. For the Intel you have to infer from field 249. To do this, divide the raw number by your disk size in GB, mine is currently 457 divided by 180 = average erase about 2 or 3 times. I purchased the SSD520 earlier than the M4 but I did not install Win7 on it (as well as the fact it's larger), thus a bit "younger" than my M4.
I could not discern this data from the SMART data from the 32G disk on my eeePC however. Its smaller size may reduce its life however.
(Other than the fact I'm using a tmpfs to build I don't do anything special to reduce writes to the disk. Both my M4 and SSD520's endurance loss due to normal use have been going down *really* slowly, slower than what one would normally think...) _________________ Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
Back to top |
|
|
|