Why did I bother

[Nec_V20]

Virge closed down the previous thread with the comment - which I cannot reply to:

I left this thread open in hopes of seeing some meaningful and insightful discussions. That has not happened. So this thread is closed.

On a side note, Nec_V20, we are (generally speaking) a "proof positive" forum. I.E. as BrightCandle puts it, "He who asserts must prove". These are after all the AnandTech forums. So go be like Anand and do some science. If you had provided rigorous and sound benchmarks alongside your commentary, this thread would have been far better for everyone involved. As it stands however you're making a somewhat wild claim (60%) without the necessary hard data to back it up.

-ViRGE

So I will answer him here instead of in a private message which I am precluded from sending.

The insightful post occurred in the first one in the thread, after that the thread was infested by cognitive flatulence.

Within the context of the original post the concept of a "benchmark" is meaningless.

The benchmark with regard to my own system (Intel core i7 990x) and C&C Zero Hour would be 0%.

However in the context of my neighbour's laptop where he would click his mouse and it took seconds to react at the lowest possible graphics settings but still in the maximum resolution of his screen whereas with the fix the response of the game was immediate, what percentage would you apply to that? How many hundreds of percent?

It is the example that came to mind because it is the crassest difference I have personally experienced.

I would never have joined Anandtech or posted my original post here if not for something someone wrote to me:

That's a good tip. You should post it on Anandtech or Tomshardware or someplace where it could help more people.

And as you may have noticed I had no problem whatsoever in answering any and all queries, no matter how ignorantly or facetiously stated because I was not voicing an opinion, but imparting factual knowledge - after being a computer techie for 30 years I do know the difference.







What don't you understand about a mod locking your thread.

You do not get to create another and pickup where you left off.

Now leave it alone.


If you have an issue with the moderation, create an MD thread with the link to your left.

Do not create a thread here to post your rebuttal about the moderation.


esquared
Anandtech Forum Director

 

[Vesku]

Well, there is some historical "evidence" regarding CPU affinity. I remembered when World of Warcraft first went multi-thread setting affinity helped quite a bit:

http://www.mmo-champion.com/threads...ecture-and-WoW-Lag-and-Graphical-tweaks-Guide

I'm doubtful of its benefits in modern software though, current World of Warcraft seems to do fine without user tweaking of affinity.

Edit: Haven't heard the theory regarding older single thread software, affinity, and caching. Perhaps you could compile a list of links to similar suggestions of setting affinity (as there was for early multi-thread WoW) for such games.

 

[Nec_V20]

Well, there is some historical "evidence" regarding CPU affinity. I remembered when World of Warcraft first went multi-thread setting affinity helped quite a bit:

http://www.mmo-champion.com/threads...ecture-and-WoW-Lag-and-Graphical-tweaks-Guide

I'm doubtful of its benefits in modern software though, current World of Warcraft seems to do fine without user tweaking of affinity.

Edit: Haven't heard the theory regarding older single thread software, affinity, and caching. Perhaps you could compile a list of links to similar suggestions of setting affinity (as there was for early multi-thread WoW) for such games.

With regard to X3 it did cure a number of CTDs (Crash To Desktop) - including my own - problems when the X3 engine ran on its own core and all other processes were precluded from it.

The reason for my post was expressly because there is an utter dearth of public sources for this piece of arcane knowledge.

I would really like to know how I could have made my original post any clearer.

I was not making a post about some vague, random setting of affinity and how it might hopefully help but rather EXACTLY HOW to set affinity so that it WILL help.

 

[Maximilian]

I honestly didn't see a problem with your original thread OP but yeah some evidence would be nice to see.

I don't know too much about scheduling etc, ive only just recently delved into it in my computing degree but a context switch (think that's the right term?) moving the process to another core arbitrarily does seem like it would decrease performance but by how much? Well that's why we need benches.

Anyways from what ive seen with single threaded games, the windows scheduler may have been updated to deal with them better making messing with affinity pointless. I remember seeing 25% overall CPU utilization in task manager spread across 4 cores (possibly on vista years ago) whereas now i see the same 25% overall but only on 1 core (windows 8). Pretty sure this was with sins of a solar empire, a single threaded game that runs like a one legged turtle stuck in tar on a large map with many AI's.

 

[Nec_V20]

I honestly didn't see a problem with your original thread OP but yeah some evidence would be nice to see.

I don't know too much about scheduling etc, ive only just recently delved into it in my computing degree but a context switch (think that's the right term?) moving the process to another core arbitrarily does seem like it would decrease performance but by how much? Well that's why we need benches.

Anyways from what ive seen with single threaded games, the windows scheduler may have been updated to deal with them better making messing with affinity pointless. I remember seeing 25% overall CPU utilization in task manager spread across 4 cores (possibly on vista years ago) whereas now i see the same 25% overall but only on 1 core (windows 8). Pretty sure this was with sins of a solar empire, a single threaded game that runs like a one legged turtle stuck in tar on a large map with many AI's.

Name a benchmark which would measure it. I know a lot of them but none that would measure what I was discussing. With regard to CTDs I suppose there is a benchmark, it is called the Event Viewer

 

[BUnit1701]

Name a benchmark which would measure it. I know a lot of them but none that would measure what I was discussing. With regard to CTDs I suppose there is a benchmark, it is called the Event Viewer

It would seem if you cant measure it, its not there...

Take underperforming game, meausre FPS, core park said game, measure FPS, show difference. Its not complicated.