搜尋 Mozilla 技術支援網站

防止技術支援詐騙。我們絕對不會要求您撥打電話或發送簡訊,或是提供個人資訊。請用「回報濫用」功能回報可疑的行為。

Learn More

How to get hardware acceleration working (on somewhat older laptop)

  • 9 回覆
  • 5 有這個問題
  • 4 次檢視
  • 最近回覆由 user633449

more options

I have a somewhat older laptop running an Nvidia-based GeForece 440 Go 64 graphics card.

Using about:support to analyze FF's performance, it shows hardware acceleration was "0/1. Blocked for your graphics driver version."

What can I do to get this acceleration working? There are no more recent driver versions available. Will this be resolved with older versions of FireFox? (I'm only running 10 on this machine, anyway, as I never did buy into the whole "rapid release" approach)

I have a somewhat older laptop running an Nvidia-based GeForece 440 Go 64 graphics card. Using about:support to analyze FF's performance, it shows hardware acceleration was "0/1. Blocked for your graphics driver version." What can I do to get this acceleration working? There are no more recent driver versions available. Will this be resolved with older versions of FireFox? (I'm only running 10 on this machine, anyway, as I never did buy into the whole "rapid release" approach)

所有回覆 (9)

more options

To get the best performance from Firefox (plus security and performance issues) you should update to Firefox 21 IMMEDIATELY (I can't stress this enough).

Second, HWA is often blocked for older cards and drivers for stability reasons. circumventing these blocks isn't supported.

more options

When you say "isn't supported", shall I presume there's a way to do it anyway, supported or not?

After all, I'd rather have the choice to sometimes have a less-stable but high-performing laptop than one which crawls around on some sites such that it's pretty useless.

Why do you stress, so strongly, updating to the latest version? After all, it seems the performance issues I have are with getting the hardware acceleration to work. I've lost a lot of faith with the rapid release program as it stresses "always updating" as opposed to updating when it makes sense. I've been around software long enough to know that it's often best to wait until a version is PROVEN before updating, and sticking to what is known to work until then.

more options

Why have you lost faith in Rapid Release? modern browsers are migrating away from this thought process you seem to have, as a new version of Firefox is released, the old one is deprecated (same with Chrome, IE is migrating to this process, Opera, etc.) For example, as soon as Firefox 21 came out, Firefox 20 was no longer supported, the security issues fixed in Firefox 21 were disclosed (meaning that Firefox 20 has public security issues now), and no updates will ever be released for Firefox 20. We watch Firefox 21 feedback, and there are no bugs in it, so we aren't going to release any bugfixes for Firefox 21 until Firefox 22 is released, at which point Firefox 21 will be deprecated. Staying on older versions means you have known and publicly disclosed security holes, and you are running a version of Firefox that has known bugs, crashes, stability issues and performance issues that have been fixed in future versions. Staying on old versions isn't an option anymore (it never really was). You don't have to update the day of release (even with auto-updates on, Firefox won't update the day of release anyway, we wait several days after each release before turning that on for all users, (this is called throttling). If thee is a major issue in a release, we will release an update before many users get the bad update.

And no, there is no way to force Hardware acceleration (and updating to a new version of Firefox will give you better performance than forcing HWA anyway, especially on such old hardware). You can do weird things like spoof your graphics card, but the chances of breaking Firefox quite badly aren't worth it.

more options

Well, we could discuss rapid release for a long time but suffice it to say I believe version numbers should MEAN something. They should also tell us a little about the significance of the updates. With rapid release, it's nothing more than a number corresponding to a date. You could go 5 version numbers without a noticeable update, or have a tremendous one each release for those 5. As a person who has done publishing and editing, I can't imagine the same significance being put to a revision where I caught 2 minor spelling errors as there is to a revision where I deleted a chapter, added a new one and rewrote yet another.

Simply put, it makes me have less faith in the priorities of the direction of development. All the rhetoric of "...away from the mindset you seem to have" doesn't mean that this mindset is actually such a bad thing.

Anyway, no matter. I just want to address this performance issue. It seems hard to believe a new FF version could increase speed above that of hardware acceleration. As a person who also has done 3D graphics, I know that you could have a newer and better piece of software but nothing could beat even the least, oldest amount of hardware acceleration. The world before DirectX and OpenGL was a slow one, indeed.

What I would prefer is the OPTION of balancing instability with acceleration. Sometimes manually switching between the two depending what I want to do that day.

I did a little poking around and found some settings to force WebGL and direct2D. I didn't find out how the integer settings in the following, though, affect performance and the ability to override the forced blacklist:

gfx.blacklist.layers.direct3d9

more options

I'm not going to debate with you, but your perception of Rapid Release is wrong. Firefox 21 is substantially faster than Firefox 10 (try it yourself or look at any benchmarks), forcing graphics acceleration on unsupported hardware is going to leave you with nasty bugs (we don't just block graphics cards and drivers willy-nilly, we block them for a reason, bugs), version numbers are becoming irrelevant in the browser game anymore (who cares if there was a major change in version X or not, update if the version you are running is unsupported). Have you even tried Firefox 21?

more options

The nVidia GeForce4 440 Go according to here is from early 2002.

https://wiki.mozilla.org/Blocklisting/Blocked_Graphics_Drivers

more options

I agree, it's best not to turn it into a huge discussion (which I'm sure has been batted around elsewhere, before). I just want to point out that I didn't mean to imply whether a certain version (e.g. 21) is good or bad. 21 could, indeed, be the best thing ever. I just find that when updating, especially on older equipment, like that laptop, it can invite a lot of problems. Some new versions have horrible memory management (I seem to recall 7 was a huge spike) or other issues. The race to be the most current has to be, sometimes, balanced with making a version which is rock-solid, but with newer versions available for users who want the cutting edge. It doesn't mean that version #s are irrelevant, it's just whether you care if they mean anything or not.

Do you have any info on version comparisons and known issues? I'm perfectly willing to look into things to see which version is worth trying, and whether that's the latest. I just want to know what the issues are before having the headache of having to revert if there's a problem.

I understand that if bugs are serious enough, a driver would be blacklisted. I would just hope that, when adhering to a standard, such bugs would only be "so" bad.

more options

There aren't versions to compare, that's the issue. Firefox 20 replaced Firefox 19, and Firefox 21 replaced Firefox 20. There aren't options, "I want to use 19, or 20" it's "I used 19, now I use 20, and now I use 21". You can read release notes with both new features and improvements, along with known issues at http://www.mozilla.org/en-US/firefox/21.0/releasenotes/ (change the version for each version you want to look at). But regardless, you need to start updating.

more options

As for what we decide to blocklist, yes, we blocklist for bad bugs, startup crashes, unreadable text, major performance decreases, etc.