5 ms·

thinkpad T420 user here running archlinux. i agree with none of these points, actually, except #1, which i wouldn't know about because i'm not using dual-mode

thinkpad T420 user here running archlinux. i agree with none of these points, actually, except #1, which i wouldn't know about because i'm not using dual-mode graphics card. "2. inability to use multiple monitors" < incorrect. i use a docking station with a dell u2711 hooked up to it. pressing fn+f7 switches correctly between external, internal, and external+internal with no problems. "3. some ports being unusable" < incorrect. every port on my thinkpad is usable, including the ports on the docking station. "4. audio glitches" < incorrect. my audio works better with my u2711 (via display-port audio) than it does over my macbook air (via display-port audio). why? because i can control the volume via the volume control via linux+pulse whereas OSX artificially restricts me -- it tells me i can't control the audio levels via mini-display port unless i'm using an apple display. "5. hand patching required to keep the laptop stable" < incorrect? by this, do you mean updates? don't all OSs benefit from updates? if you mean configuration, udev (now systemd-tools) found everything correctly on my thinkpad on the first go. this was the case when i recently ran ubuntu 12.04 off of a flash drive as well. "6. poor support for X11 even with simple intel integrated graphics" < incorrect. i have the intel HD3000 on my thinkpad, which runs everything pretty much the same way that my macbook air (with the intel3000) runs them. flash video at fullscreen with no tearing, etc (in fact, flash runs better on linux than it does on my macbook). regarding battery life, i did have to do a bit of tinkering there. but that's because i'm a bit power-consumption obsessed. i used powertop to show me what needed fine tuning, then wrote a power management script that changes my computer's parameters depending on whether it's plugged in. the result: an idle 7w power consumption (with wi-fi on). that's roughly the same power consumption i get from the macbook air, which is an ULV chipset. battery life for windows on my T420 is ~10 hrs. with linux,...~12hrs. "no competitively priced linux laptops"? my T420 was $1000 with a 2520M processor, 1600x900 resolution display, 160GB intel SATA 3 SSD, and 8GB of RAM.

nonviable 14y ago
Exactly. My primary device is a daily updated (with aptitude) Debian testing on a T410 and have exactly none of the specified problems, though I too don't run a dual-mode graphics card. Not sure where the parent post got their info. Lenovo took a big step down from IBM when they took over the brand, but for my money they are still the best non-mac laptops out there for a Linux user.
- KirinDave 14y ago
  If they are "best" is immaterial. Linux is a poor laptop experience. And you don't have a dual mode graphics card and probably only have VGA out, so you really don't have dual monitor support except in entirely unsatisfactory or projector contexts. You also almost assuredly get battery life that is worse to a comparable windows 7 installation with Lenovo customizations.
  - pyre 14y ago
    You also almost assuredly get battery life that is worse to a comparable windows 7 installation with Lenovo customizations. Why is this though? This has been the case for a while, but I'm unclear as to what optimization happens on Windows/OSX that doesn't happen on Linux. It's usually attributed to the Windows/OSX manufacturer drivers, but what are those drivers doing differently? Anyone have even a general idea?
    - SkyMarshal 14y ago
      The most recent Linux kernels make big improvements there. My linux laptop battery life went from ~4hrs to ~8hrs when I upgraded from Ubuntu 11.10 to 12.04. The reason is a few patches added to the kernel in 2012: http://www.h-online.com/open/news/item/New-Linux-kernel-fixes-power-saving-issues-1429482.html http://www.h-online.com/open/news/item/New-Linux-kernel-fixe... That article is from Feb 2012 and mentions two big power-saving problems with the kernel, the first one patched, the second (RC6 power-saving feature in Sandy Bridge) unpatched as of the article's publish date. The second has since been patched as well, for huge power gain.
      - KirinDave 14y ago
        By the way, I agree that recently things can get better. I stopped trying to get the Linux laptops in Jan after about 5 months of wrangling.
KirinDave 14y ago
Even with powertop and and eye for cutting things, odds are you are ending up with worse battery performance than a Lenovo-supported drop of Windows 7. Which is sad because it's not great battery life for not great performance, the T420 is also a pretty underpowered machine for developer use. I'm not going to fisk your post otherwise. I have better things to do. I will point out that due to the way Lenovo updates models over time, it's entirely possible for mutiple people to buy "T420s" and end up with different-enough components that you'd need to tweak drivers to get things to work right. So "Your T420" may not be someone else's "T420." This is provably the case for my T520, and supposedly the case for the T61.
- eitally 14y ago
  Actually, you're mostly wrong on that last point. We are an enterprise customer of Lenovo's for both laptops & desktops and they are quite good about maintaining stability within a model range for at least 9-12mo, and they always tell us at least a quarter in advance of any changes... and they provide imaging services if a customer wants to pay for it (Windows or Linux). There's still enough IBM alive at Lenovo for them to care about engineers.
  - KirinDave 14y ago
    Their consumer facing division cared little for me or my problems. Evidently I didn't pay enough for support, even though my problems were all well within the first 90 days of purchase. I've had some of the worst customer service I can imagine from them, both at the policy and individual level. But it is ABSOLUTELY the case that my T520 has a slightly different configuration than most of the T520s I come across. There is no argument allowed here: this is fact. It's also ABSOLUTELY the case that Lenovo couldn't replace my latpop with a new one because I ordered a very popular processor option that was repeatedly out of stock. Twice now I've been offered replacement downgrades without even so much as an offer of credit. Twice I've refused, and twice I've gotten back a laptop with the same exact problem. So perhaps I am just not enterprisey enough.
SkyMarshal 14y ago
> the result: an idle 7w power consumption (with wi-fi on). that's roughly the same power consumption i get from the macbook air, which is an ULV chipset. battery life for windows on my T420 is ~10 hrs. with linux,...~12hrs. That's pretty sweet. Any chance you could post that script on Github? I've been tweaking my linux laptop power settings too and would love to take a look at it.
- kamechan 14y ago
  Ahh, yes. Thanks for asking. It occurred to me I probably should have included this info in my original post only after the window for editing it had passed. I'd previously written a bit about how to optimize the Thinkpad power consumption on another HN post here: http://news.ycombinator.com/item?id=3564400 http://news.ycombinator.com/item?id=3564400 FYI, I find that the kernel flags are necessary even with the latest kernel. I'm currently on 3.3.8-1-ARCH and experimented with turning them off because the power regression that started in kernels > 2.6 was supposed to have been fixed and merged back into the kernel code. But I found that passing the pcie_aspm=force flag at boot time still results in a significant savings. Indicating that the regression (for at least my configuration) might still be there. If you have an Intel video card, then check out the other kernel args I'm passing as well for additional power savings (and there are more parameters you can play with outlined here http://www.phoronix.com/scan.php?page=article&item=intel_i915_power&num=1 http://www.phoronix.com/scan.php?page=article&item=intel...) I experimented with a few different combinations before settling on the ones I posted in my HN link, but they're what gives me optimal battery life on my T420. Despite what the other poster had said, I'm getting battery life now that's significantly better than what I get booting into Windows on the same machine, which usually stays at around 9-10 watts idle. I'm getting 6.8-7.8W Without any optimization, my battery life was much less than I'd expected to get from the 9-cell (~7 hours). But with the power management script and the kernel flags, I can get 13 hours with light use and 10-11 while doing development (i.e. apache, mysqld, eclipse, multiple browsers, etc... all running). edit: FYI powertop2 was the beta release of powertop at the time. the source has been merged back into powertop so you can just use that.

nonviable 14y ago

Exactly. My primary device is a daily updated (with aptitude) Debian testing on a T410 and have exactly none of the specified problems, though I too don't run a dual-mode graphics card. Not sure where the parent post got their info. Lenovo took a big step down from IBM when they took over the brand, but for my money they are still the best non-mac laptops out there for a Linux user.

KirinDave 14y ago
If they are "best" is immaterial. Linux is a poor laptop experience. And you don't have a dual mode graphics card and probably only have VGA out, so you really don't have dual monitor support except in entirely unsatisfactory or projector contexts. You also almost assuredly get battery life that is worse to a comparable windows 7 installation with Lenovo customizations.
- pyre 14y ago
  You also almost assuredly get battery life that is worse to a comparable windows 7 installation with Lenovo customizations. Why is this though? This has been the case for a while, but I'm unclear as to what optimization happens on Windows/OSX that doesn't happen on Linux. It's usually attributed to the Windows/OSX manufacturer drivers, but what are those drivers doing differently? Anyone have even a general idea?
  - SkyMarshal 14y ago
    The most recent Linux kernels make big improvements there. My linux laptop battery life went from ~4hrs to ~8hrs when I upgraded from Ubuntu 11.10 to 12.04. The reason is a few patches added to the kernel in 2012: http://www.h-online.com/open/news/item/New-Linux-kernel-fixes-power-saving-issues-1429482.html http://www.h-online.com/open/news/item/New-Linux-kernel-fixe... That article is from Feb 2012 and mentions two big power-saving problems with the kernel, the first one patched, the second (RC6 power-saving feature in Sandy Bridge) unpatched as of the article's publish date. The second has since been patched as well, for huge power gain.
    - KirinDave 14y ago
      By the way, I agree that recently things can get better. I stopped trying to get the Linux laptops in Jan after about 5 months of wrangling.

KirinDave 14y ago

Even with powertop and and eye for cutting things, odds are you are ending up with worse battery performance than a Lenovo-supported drop of Windows 7. Which is sad because it's not great battery life for not great performance, the T420 is also a pretty underpowered machine for developer use. I'm not going to fisk your post otherwise. I have better things to do. I will point out that due to the way Lenovo updates models over time, it's entirely possible for mutiple people to buy "T420s" and end up with different-enough components that you'd need to tweak drivers to get things to work right. So "Your T420" may not be someone else's "T420." This is provably the case for my T520, and supposedly the case for the T61.

eitally 14y ago
Actually, you're mostly wrong on that last point. We are an enterprise customer of Lenovo's for both laptops & desktops and they are quite good about maintaining stability within a model range for at least 9-12mo, and they always tell us at least a quarter in advance of any changes... and they provide imaging services if a customer wants to pay for it (Windows or Linux). There's still enough IBM alive at Lenovo for them to care about engineers.
- KirinDave 14y ago
  Their consumer facing division cared little for me or my problems. Evidently I didn't pay enough for support, even though my problems were all well within the first 90 days of purchase. I've had some of the worst customer service I can imagine from them, both at the policy and individual level. But it is ABSOLUTELY the case that my T520 has a slightly different configuration than most of the T520s I come across. There is no argument allowed here: this is fact. It's also ABSOLUTELY the case that Lenovo couldn't replace my latpop with a new one because I ordered a very popular processor option that was repeatedly out of stock. Twice now I've been offered replacement downgrades without even so much as an offer of credit. Twice I've refused, and twice I've gotten back a laptop with the same exact problem. So perhaps I am just not enterprisey enough.

SkyMarshal 14y ago

> the result: an idle 7w power consumption (with wi-fi on). that's roughly the same power consumption i get from the macbook air, which is an ULV chipset. battery life for windows on my T420 is ~10 hrs. with linux,...~12hrs. That's pretty sweet. Any chance you could post that script on Github? I've been tweaking my linux laptop power settings too and would love to take a look at it.

kamechan 14y ago
Ahh, yes. Thanks for asking. It occurred to me I probably should have included this info in my original post only after the window for editing it had passed. I'd previously written a bit about how to optimize the Thinkpad power consumption on another HN post here: http://news.ycombinator.com/item?id=3564400 http://news.ycombinator.com/item?id=3564400 FYI, I find that the kernel flags are necessary even with the latest kernel. I'm currently on 3.3.8-1-ARCH and experimented with turning them off because the power regression that started in kernels > 2.6 was supposed to have been fixed and merged back into the kernel code. But I found that passing the pcie_aspm=force flag at boot time still results in a significant savings. Indicating that the regression (for at least my configuration) might still be there. If you have an Intel video card, then check out the other kernel args I'm passing as well for additional power savings (and there are more parameters you can play with outlined here http://www.phoronix.com/scan.php?page=article&item=intel_i915_power&num=1 http://www.phoronix.com/scan.php?page=article&item=intel...) I experimented with a few different combinations before settling on the ones I posted in my HN link, but they're what gives me optimal battery life on my T420. Despite what the other poster had said, I'm getting battery life now that's significantly better than what I get booting into Windows on the same machine, which usually stays at around 9-10 watts idle. I'm getting 6.8-7.8W Without any optimization, my battery life was much less than I'd expected to get from the 9-cell (~7 hours). But with the power management script and the kernel flags, I can get 13 hours with light use and 10-11 while doing development (i.e. apache, mysqld, eclipse, multiple browsers, etc... all running). edit: FYI powertop2 was the beta release of powertop at the time. the source has been merged back into powertop so you can just use that.