NVIDIA driver slowdowns

[spydrct02]

OK, explain why we should all restrict ourselves to 50 MHz processors that aren't broken.

Yes, the drivers for other OSs seem to work better. Maybe just an XP/ForceWare 32 bit driver problem? Even Vista ran fine with all 60+ services running, though a little slower than Linux, yet more consistent. In Linux, any hard drive activity would slow Graal down, but level changes were absolutely instant every time.

Still, before you say your Graal runs fast, try doing those tests in maxed fullscreen like in the two videos I've made. Good luck on not getting a failed, though, if you enter and exit too fast.

I ran Graal in 2048 x 2048 (highest it supports) today and discovered that the level change times were one full second on 66.00 ForceWare drivers. If I were to run on the same resolution on 90.00 drivers, that would likely be an easy 5 second lag on level changes. I'll eventually test that resolution on the newer drivers. There's at least one good thing I've found from that, though - throwing snowballs at people from 2 levels away can be fun.

Originally Posted by ReBorn_Spirit

Well, now that my Graal is running on 64 bit XP
**Seems that the DirectDraw Graal uses has effect on Windows XP x64, slowing down a few animations here and there, also graal will slow down when too many people are around, or npc's

Probably because DirectDraw is using all of your processor up. If there was an OpenGL mode, it'd use more video card and free up the processor, resulting in less/no lag.

[ReBorn_Spirit]

Originally Posted by spydrct02

OK, explain why we should all restrict ourselves to 50 MHz processors that aren't broken.

Yes, the drivers for other OSs seem to work better. Maybe just an XP/ForceWare 32 bit driver problem? Even Vista ran fine with all 60+ services running, though a little slower than Linux, yet more consistent. In Linux, any hard drive activity would slow Graal down, but level changes were absolutely instant every time.

Still, before you say your Graal runs fast, try doing those tests in maxed fullscreen like in the two videos I've made. Good luck on not getting a failed, though, if you enter and exit too fast.

I ran Graal in 2048 x 2048 (highest it supports) today and discovered that the level change times were one full second on 66.00 ForceWare drivers. If I were to run on the same resolution on 90.00 drivers, that would likely be an easy 5 second lag on level changes. I'll eventually test that resolution on the newer drivers. There's at least one good thing I've found from that, though - throwing snowballs at people from 2 levels away can be fun.


Probably because DirectDraw is using all of your processor up. If there was an OpenGL mode, it'd use more video card and free up the processor, resulting in less/no lag.

Which processor? 0 or 1?

[spydrct02]

In the processes tab (task manager), it should only show the total CPU usage Graal is taking, and not which "processor" it's going to. If not, set the affinity to either one and check.

[ReBorn_Spirit]

lol
You can set your affinity to either have the program run on processor 0 or 1
or both.

[spydrct02]

Yes, but following the total CPU usage through the performance tab's graphs isn't reliable for tracking one process, is it? To see the usage split through both cores, you have to be in the performance tab with the graphs. You want Graal's total usage, not split CPU usage.

Or do you mean there's a separate CPU percentage per core in the processes tab? I'm pretty sure the processes tab only shows the total amount the program is using. I could be wrong, though, I've not had a real chance to screw around with multi-core machines besides through VNC.

Maybe next time point out where in my post you were responding, so I know what you're actually asking. I was lost since the start, and this is the result.

I think by "which processor," you mean which core, but are you aware you only have one physical processor? Task manager shows two graphs, but both added up equal your processor. I said set Graal to one affinity so you don't have to add both up and get unreliable results in the graphs from other programs taking CPU usage too, but I said to do that IF you can't get the process' (Graal in this case) total usage reading some reason.

[ReBorn_Spirit]

Clarification for you. Windows and many other programs can run on core 0, while graal could run on its own on core 1 for an example. Now if graal locked up and used 100% of core 1. Why would I care now? I could use all of core 0 still to access all my programs and do whatever I so desired, because I still have a processor to use that isnt busy.

Setting Graal to a core 1 with affinity like that could easily solve any windows lagg I could possibly have.

And yes, it does have seperate CPU usage PER core. Meaning it would show up 2 graphs instead of just 1 total graph. To use Dual core or Dual Processors you need XP Profession or XP Professional x64. Because the Windows XP Home edition will not utilize more then 1 core at any time. Even if you had 3 or 4, it still would use only one with XP Home. That by the way, is something Microsoft doesn't try to tell you.

[spydrct02]

This whole thing is only half-relevant to the whole topic. You're discussing how Hyperthreading/dual cores work, and I was originally talking about total processor usage. That's assuming your question was a reply to the last bit of my post about the DirectDraw processor usage, anyway. You have yet to actually tell me what you were even replying to.

When I said it's using all of your processor up, that means your processor, not one of the cores. Sure, you could estimate both cores' CPU usage, but it's much easier and reliable going by the CPU usage that's said in the processes tab. That is the CPU usage I was mentioning, the one in the processes tab - the actual total CPU usage, not the split CPU usage. Where you're confused is you're thinking each graph represents a processor. Those graphs are your two cores; both cores equal your processor.

If you still don't get what I'm saying, just take "both" as an answer to your question. If you have it set to a single affinity, then the there is no need for the question. Either way, we're getting nowhere by this whole thing. Maybe you should try my tests and post your results so you have something to compare with everybody else, instead of saying something subjective like, "it runs great."