Talk:Java Edition hardware performance

About the "Sapphire OC Edition" and 850 MHZ
Hey, I posted the i5 2500 and the Sapphire Radeon 7750 OC Edition 1GB GDDR5. It got changed to "AMD" but Sapphire might have put something like a better fan in the OC Edition of their products. Plus, the average framerate is gone. Please add them back for the sake of readers and shoppers.


 * The proper product title is listed in the GPU column. If you wish, you may note the GPU's sub-manufacture in the comments. All that matters for the GPU column is core frequency and RAM amount (and frequency if you wish to to so). Temperatures related to overclocking/underclocking or fans and their speeds have no bearing on these performance results. Also, for FPS results you would either note the avg. FPS or the FPS range. Range is preferred, and you have provided it as such. -- Bb 20 (Edits)  17:31, 17 January 2014 (UTC)

A note for 'fps_window' and 'fps_window_native'
'fps_window' is the window resolution that Minecraft usual starts up as, which is 854×480. This use to be the window in it's maximized state. 'fps_window_native' is the same old 'fps_window', though is now defined as the primary native monitor resolution.

The reasoning for this new column is to provide more details for Minecraft users to compare against. Right now, users have to compare against a GPU + monitor resolution to find what best represents them. Now users should be able to compare against GPU + 854×480 and GPU + + monitor resolution. I hope that makes sense. So now when I want to see what works for me I am not bound by comparing to a screen resolution I may not have. While we could have users benchmark at common resolutions such as 1,280×720, 1366×768, 1920×1080, 2560×1440/1600 for more comparison data, it's easier to use one that everyone should be able to run at, and then their native monitor resolution.

I will be updating this page further to reflect these changes. -- Bb 20 (Edits)

Sort framerate columns by data-type
This is something I've wanted to do for a while, though not sure how to do, if doable at all with multiple columns; sort all framerate columns using data-sort-type="number". If someone could lend me help on how that is done then that would be most helpful! -- Bb 20 (Edits)  17:33, 6 September 2014 (UTC)

Mac issues
There are two big issues for mac users:


 * 1) We don't have fraps to tell us the frame rate, only the native F3 screen.
 * 2) Maximized does not actually maximize. We have that menu bar on the top, so even with the dock removed, we still have a slight screen ratio error.

Keybounce (talk) 19:27, 6 September 2014 (UTC)


 * It never says you must use fraps, it says it is optional. Just use the native F3 screen.
 * Take the results you get from that then
 * --KnightMiner  (t 19:30, 6 September 2014 (UTC)


 * What Knight mentioned. Also, for verification, does OS X currently maximize windows while keeping their current window ratio? In the latest OS X beta, I believe, maximizing a window definitely extended the borders to the screen edge for a true maximization. If that verification if true then I'll need to throw in an extra instruction for OS X users. -- Bb 20 (Edits)  23:47, 6 September 2014 (UTC)

Verification of world?
Is there a way to verify what we are supposed to see at 276, 82, 180?

I'm in the middle of a thick forest, with uneven ground -- there is no way to make a small circle without the terrain around me affecting what happens.

Keybounce (talk) 19:26, 6 September 2014 (UTC)


 * As long as you are using the current version stated in the instructions and go to the exact coordinates then you won't have to worry about it. We all use the same world gen code so the landscape won't change. Also, I've been thinking of how to accurately, and painlessly, capture FPS - this includes walking in a circle. That's been in the instructions for quite the while actually, though looking in all directions in spot may be better to do than walk around. Otherwise, you may just want to walk around obstacles to complete your circle. -- Bb 20 (Edits)  23:39, 6 September 2014 (UTC)


 * Also, the screen resolutions you used for both AMD Radeon tests, are they native resolutions? Only the first one looks like it is; 1440×900. -- Bb 20 (Edits)  23:57, 6 September 2014 (UTC)

Catalyst UnifL
A revision was made relating to software (Catalyst UnifL) for OEM hardware; revision linked here.

Could a user clarify a few points for me, please? From the lack of information I got from leshcatlabs, UnifL installs non-OEM drivers /based on/ AMD & Intel drivers along with custom Catalyst software; to help assist with "Switchable Graphics". AFAIK, this is something that is already supported in AMD's Catalyst software - CCC > Power > Switchable Graphics. Another thing this software does is keep the OEM hardware up-to-date, something that the instructions in the article point out how to do if the user doesn't do this already. -- Bb 20 (Edits)  07:41, 17 November 2014 (UTC)


 * For the sake of consistency, I think we should have people use the official OEM video drivers, not some third-party one that has been modified in unspecified ways. -- Orthotopetalk 09:02, 17 November 2014 (UTC)


 * Wanted to make sure I wouldn't run into a conflict somewhere along the line for it, though I do agree that consistency is for the best. Thanks for your input! -- Bb 20 (Edits)  09:13, 17 November 2014 (UTC)

Wrong driver version listed in post
When I made my performance entry, my comment was "using latest Intel drivers". Bb 20, the main "mod" of this page, added the driver version number (15.36.18.4156). I had meant 10.18.14.4080, as that was the version I benchmarked on and the version I believed to be the latest, but apparently 4156 had been released on March 6, and I did not know about this. I tested performance on 4156, and it seems worse than on 4080 (it also broke SEUS support as well, but that's another story). How do I go about editing my entry? --Wwtoonlinkfan (talk) 03:03, 6 April 2015 (UTC)


 * You should be able to simply edit the section and correct the driver number. (Make sure to add an edit summary stating the reason, as you did not add your username to the entry title.) – KnightMiner  t/c 03:15, 6 April 2015 (UTC)


 * My mistake, though I'm glad you pointed that out! Good note for SEUS too. I pulled the latest driver release version number from here if you're curious. I do my best to insert missing information with what's available. However, sometimes other users have to correct us, you know? -- Bb 20 (Edits)


 * It's ok - honestly, it's more my fault for not checking the driver versions. Thanks for the correction! --Wwtoonlinkfan (talk) 05:58, 6 April 2015 (UTC)

Changing the format to add the entries slightly
Since the current format basically prevents the ability to have categories in the right place and have a navbox, why not move the entries additions to a subpage and just transclude it here? The "Add entry" link can easily be pointed there, plus it would allow us to easily hide the unfiltered results until they are filtered. – KnightMiner  t/c 03:21, 6 September 2015 (UTC)


 * I'm not familiar with transclusion's function in its totality, however, I feel that would be a good workaround. -- Bb 20 (Edits)


 * Done. The entries are now stored on Hardware performance/entries. I've set it up so the unfiltered results are by default not shown (so they can be reviewed and sorted first), but feel free to change that if you prefer users to see them. – KnightMiner  t/c 00:44, 10 September 2015 (UTC)

(Required > Optional) Drive requirement for benchmarking
Gave it some thought, ran a series of benchmarks, and my results seem to indicate that the storage drive has little to no impact on frame rates. The motivator: many people forget, don't bother, cannot find, or  and as a result the drive model is missing from their entry which means that the entry itself may be scrapped due to incompleteness. As such, to keep more entries from being undone so we have a larger sampling(?) of hardware to compare with, I'm going to change this to be optional.

Benchmarks from my test can be found here - forgive any wrong terminology or technical inaccuracies. This was done on my Acer Aspire v3-571G and is listed as an entry as well. -- Bb 20 (Edits)  05:45, 9 March 2016 (UTC)