OvermindDL1

There is a limit, but it is quite large right now until someone does something stupid. Feel free to post, it is already created.

Ah, so the py script needs to be changed to detect windows and launch the diff directly I take it?

MCF kind of sucks, so I have no issues with many of those, probably make all three of those into child boards to reduce main screen space.

Time to do a search of one at a time.

Ah, I was reading the old information, ignore me then. If you need help with placing textures/sounds in the jar then make a post in the support forum along with code of how you access those now and we can help you to change it to work internally (if it would not already, you would be surprised).

This. And do not use ModLoader.addOverride, that is old, deprecated, and rather useless now, use a new texture sheet instead with setTexture on the blocks/items. And it is better to use addTranslation instead of addName.

Ew, everything should be unified in py's...

Also, there should still be a stack trace at least on the terminal if you wait long enough.

I wonder how actually changing it in MC would do, would it affect FPS at all... Should be able to change it in only one place at least... The JIT definitely has a high computation cost, those 'first' readings are high. I am curious how MathHelper is faster on these now, what else is in effect. Anyone have any ideas?

Correct, but we can know that by reading the stack trace too. The stack trace is needed for pretty much all troubleshooting.

TrueBrain: If the MathHelper only gets faster on many run that might be because it gets cached, meaning Math.sin might truly be faster in the general 'random' case.

You do see the sticky thread here that says to "Read this or you may get banned"? Yes, you should read it.

You can also record exactly how you install it, showing us the resultant contents of the jar on the video as well, then running it in a terminal window and showing us what it does and prints out? And what Java version are you using?

You just listed many reasons I do not care for the JVM. If you can make a testing library for those, along with a standalone running wrapper (just a main that calls the library), along with source of course, then I can set up something on this server for your program to submit the results to. You might also want to reports what version of SSE is supported as well since it is possible that that JIT could use SSE calls to change the outcome and/or precision. Is there anything else that should be reported? Oh no derailing at all, this was actually my best hope of an outcome, proper research. On an unrelated note, I love your posts! I usually post large and detailed posts as well, but people complained to me long ago in the past so I have tried to hold back for a few years now (not always successfully), but I love the posts as they give lots of information. So no worry about holding back here, others can shorten it if necessary. EDIT1: I should probably see if I can dig up that statistical testing library in C++, if it can be ported to Java (it is not 'that' large) then it would be able to detect and bypass the jitter introduced by jitting and other things such as the cache warming up and so forth. Remind me later. And I am generally always in IRC, ping me and if I am around then I will respond, else I will respond later. Actually there might already be such a testing library for java, might try searching for one. EDIT2: Actually I think I did find one for java with a quick search, seems very similar to the C++ version I used, with a few missing features but those should not really be needed: http://www.ibm.com/developerworks/java/library/j-benchmark1/index.html The code and such is at http://www.ellipticgroup.com/html/benchmarkingArticle.html but read the article first, and he stresses hard: the only way to do pristine benchmarks is to do but a single benchmark per JVM session. And also we should have some unique identifier per computer so as if someone runs one multiple times under different load we can calculate the results from that one computer better without one person overwhelming the results. EDIT3: Also, if we get a good testing framework and reporting system set up then we could use it to benchmark other things as well, like Lex's new event system compared to MC's seemingly much worse one.

I created a new category and a few new boards for non-forge-mostly-site stuff, and since the mods board has a mostly defined aspect of how busy a mods forum should be for it to be on the front-page (~10 posts per day average for a while or more), I moved all others back to the Various Mods board, thus far that leaves Universal Electricity as the only mod on the main page, averaging well over 10 posts a day, congrats. This board will contain news items relevant to the site and possibly Minecraft in general and not to forge. If you have any requests for something to be placed here then suggest it either in a PM to me or in the off-topic board or so, but I intend to keep this boards content rather minimized for note. Created a Let's Play and other videos board for LPers and video tutorial makers and mod showcases and all else to link and advertise themselves. And there is also an Off-topic board for everything else, but please no flaming or so. Be logical or temp/perm bans can be handed out as always.

If you could convince the big name LP'rs like Direwolf etc, who make use of the Forge code, to ask their fan base to use this forum for their discussion it could very well be a boon for your team. More traffic and page views of course means more revenue for your webpage from your ads. Just to continue using Direwolf as an example, most of the posts for his thread on the minecraft forums is about getting all those mods of his to work in one instance, that is the kind of thing that should be posted and answered here I would think as they are more or less talking forge stuff. I get no ad revenue. All ads on this site help pay Lex, Eloraam, Spacetoad, and cpw (and cpw's share goes to charity), they are the ones that put the work in to forge. This server that I host the Forge site on is used for other purposes as well (although the forums have vastly become the highest use of the server), but I see not a cent. EDIT: And I created a board for these and moved this topic there. Enjoy!

That is really nice! Good job!

Oh optifine is awesome, do not be afraid to use it. Forge is designed to hook in to it.

He would know not to ask support for cracked versions, it is in the stickied post titled "Read this or you may get banned" after all. It is either something is outdated (update java), or installing it wrong.

Get diff.exe from the gnu32 utilities and put it in the right area? I use a better OS that comes with diff so I am unsure where that area is on Windows, but the error messages or code may lead you to it.

For note, that looks like an opengl error, make sure your videos drivers are up to date. And yes, do not use MCPatcher regardless.

You can always rename it from something.jar/zip to something like something.jar/zip.disabled. As long as the file extension is not jar/zip.

Try redownloading forge maybe and reinstall into MCP?

Yes indeed!

Made me curious as well, but I did not have the time to delve in to it. Also do C/C++ for a few decades here as well, I have some very nice statistical testing scaffolds that would be perfect for testing code such as this (it tests jitter of memory access, loading of various parts of memory and other things, then tests the functions based on what it found out of the cpu/memory/etc access to within a statistical accuracy, it can take a *very* long time to run, but is very accurate, found it in the boost sandbox repository years ago, made a few minor edits since). Thank you very much for the responses and your research!