6497 réponses à ce sujet

[InBleedingRapture]

Well, apparently I don't know ****, even though my overclocks are steady without any dips in PSU voltage or any system hangs. I've benchmarked extensively with 3DMark (multiple versions) and the results of overclocking my GPUs, CPU, and RAM give me about the same benches and performance as one "step above" my current gear.

My advice is not to **** around with overclocking or unlocking voltage caps unless you know what the **** you're doing, which, according to popular consensus, I do not. Since I am apparently a bull**** artist, I will refrain from trying to help figure things out when people as questions.

Modifié par InBleedingRapture, 18 mai 2012 - 03:53 .

[Dead_Meat357]

InBleedingRapture wrote...

Well, apparently I don't know ****, even though my overclocks are steady without any dips in PSU voltage or any system hangs. I've benchmarked extensively with 3DMark (multiple versions) and the results of overclocking my GPUs, CPU, and RAM give me about the same benches and performance as one "step above" my current gear.

My advice is not to **** around with overclocking or unlocking voltage caps unless you know what the **** you're doing, which, according to popular consensus, I do not. Since I am apparently a bull**** artist, I will refrain from trying to help figure things out when people as questions.

Where are you getting that from? All I said was that TDP values go out the window when overclocking. I'm not saying your video card is hitting 400watts, or that your CPU is pulling that much power. This varies wildly by system configuration. I used some of the worst case examples when I made that statement. You also don't need a 1,000 watt CPU for overclocking or anything. I never said that either. Plenty of machines, in fact most machines can operate off 600 watts or less without issue.

I wasn't saying you couldn't overclock or didn't know anything about it. But you brought up one thing I can't let go..........

3D Mark. I don't care what version it is, those results are MEANINGLESS. It's numbers are arbitrary. I am uncertain why anyone considers the numbers it generates relevant. Especially since it no longer uses an actual game engine for testing. The factors that influence it's scores are not the same factors that influence game performance. It isn't as if you can say, X,XXX 3D Marks means 94FPS in Crysis 2. It doesn't work that way. Overclocking your CPU and video card clocks helps. In some cases, tremendously so. I never said it didn't. Again there are points of diminishing returns and some factors are more important than others. Some of these factors vary by application and even more depending on the rest of the system configuration and what you are doing with it. System memory clocks are less important. They do not equal much if any game performance 95% of the time. Sure your Sandra scores go up but again, meaningless. Games are not that sensitive to memory bandwidth.

But relax man. I'm not trying to offend anyone. Just providing information. Please don't take it out of context.

Modifié par Dead_Meat357, 18 mai 2012 - 04:14 .

[neilthecellist]

Hahahaha 3D Mark... I haven't touched that program in ages.

[Dead_Meat357]

neilthecellist wrote...

Hahahaha 3D Mark... I haven't touched that program in ages.

Neither have I. I've seen people argue that it's good for stability testing. It isn't. There are much better ways to test that.

[InBleedingRapture]

3D Mark results are, in fact, not meaningless. Sure, taken by themselves, the numbers mean absolutely nothing, but the program itself does a great job of stress-testing your system. The scores are unimportant, but if I run a performance test with stock settings, then with my OC settings, then use some simple math to figure out the percentage of increase (if there is any), I'd say that it's somewhat helpful.

Obviously, it's not the only means I use for benchmarking, but it can help sort out the results. I do test on multiple versions of 3DMark as well. Prime95 is great for CPU and RAM testing and a few other programs can be useful for monitoring voltage and temps under 100% load.

Look, I'm sorry if I overreacted, but I took offense to the term "calling bull****," as if I'm just trying to arbitrarily blow hot air out of my ass. I have had no instruction in any of this, so I apologize if I make mistakes. My comments are based on a combination of experimentation, observation, and asking a few questions on certain tech boards.

[neilthecellist]

My bad, I'm sitting in accounting class, so we get very pissy easily while punching in numbers. Making a mistake means starting the f*** over at the process. I didn't mean it as a personal attack, it's just the way I usually talk when I crunch numbers on debt and equity ratios!

To bring this thread back more to discussion on textures, does anyone plan to rework the texturing on Shepard's helmet? The front part of the helmet looks really bad during cutscenes.

Also, yesterday I finished recording the game up until the Prothean Artifact mission and will be posting it on Youtube sometime later this week. Stay tuned!

[Dead_Meat357]

InBleedingRapture wrote...

3D Mark results are, in fact, not meaningless. Sure, taken by themselves, the numbers mean absolutely nothing, but the program itself does a great job of stress-testing your system. The scores are unimportant, but if I run a performance test with stock settings, then with my OC settings, then use some simple math to figure out the percentage of increase (if there is any), I'd say that it's somewhat helpful.

Obviously, it's not the only means I use for benchmarking, but it can help sort out the results. I do test on multiple versions of 3DMark as well. Prime95 is great for CPU and RAM testing and a few other programs can be useful for monitoring voltage and temps under 100% load.

Look, I'm sorry if I overreacted, but I took offense to the term "calling bull****," as if I'm just trying to arbitrarily blow hot air out of my ass. I have had no instruction in any of this, so I apologize if I make mistakes. My comments are based on a combination of experimentation, observation, and asking a few questions on certain tech boards.

Furmark and OCCT absolutely rape 3D Mark for usefulness. And increases in 3D Mark scores from overclocking one thing are still arbitrary and meaningless because they don't translate the same way into anything else. And I didn't "call bull****." I was simply trying to eliminate confusion with regard to how the hardware works and comments concerning video cards not having enough RAM, etc.

Computer hardware testing and benchmarking is one of the things I do for a living. I'm just trying to help.

Modifié par Dead_Meat357, 18 mai 2012 - 04:27 .

[InBleedingRapture]

It's no big deal. I'll admit that I am overly sensitive about (what I perceive to be) attacks against my intelligence. I need to work on that. Granted, I have been known to try to explain things I don't completely understand on occasion out of a desire to help. We're all human.

Apologies, Dead_Meat357 and neilthecellist. Nothing against you guys. I should think before posting.

And yes, I've been working on the entire N7 set just for use in my own game. It's not much of a retexture as it is just texture enhancement. I have some vids up with the armor, but not with the helmet. I'll see about capping one soon.

**EDIT**
Furmark is great. Nice and simple with reliable results. OCCT is also a cool program. I've used them both, but not for a while, since I've had my overclocks set for about a year now. Prime95 is my go-to for CPU stability testing and MEMTest is what I use for testing my memory timings.

I appreciate the help and apologize that I wasn't more receptive to it.

Modifié par InBleedingRapture, 18 mai 2012 - 04:36 .

[Dead_Meat357]

No problem. We are all human. I've had knee jerk reactions to some stuff a few times myself.

Funny thing about Prime95 is that you can run it on a Sandy Bridge based Core i3/i5/i7 CPU, while overclocked, but the second you fire up a game, CRASH city. So while I like Prime95 and do use it for CPU testing, by itself it's not always an indicator of stability alone.

[InBleedingRapture]

Yeah, the results do vary a bit with Prime95. Some modes mix in memory block transfers, and I'll admit that I'm not really sure what the difference is between testing "Short FTTs" and "Long FTTs."

Also, I'm touchy about my temps, so I don't leave it running once they start to get close to TJMax (actually the name of another fairly useful program). I've installed water cooling systems, multiple radiators, but nothing cools as well as my Zalman air cooler. Water cooling is a wast of time, for the amount of work you have to do to maintain it.

[Dead_Meat357]

InBleedingRapture wrote...

Yeah, the results do vary a bit with Prime95. Some modes mix in memory block transfers, and I'll admit that I'm not really sure what the difference is between testing "Short FTTs" and "Long FTTs."

Also, I'm touchy about my temps, so I don't leave it running once they start to get close to TJMax (actually the name of another fairly useful program). I've installed water cooling systems, multiple radiators, but nothing cools as well as my Zalman air cooler. Water cooling is a wast of time, for the amount of work you have to do to maintain it.

Good water cooling makes all the difference. Trust me. I've had all the best air coolers, and none of them can handle the heat my system puts out. If you had good water cooling hardware and it didn't work well, there was something wrong. Believe me. Using my Koolance Exos 2.5, I'm keeping my Core i7 980X @ 4.4GHz at around 37c idle. About 65c max under full load with crappy ambient air temperatures. With air cooling or the Corsair H100 my temps were 20c higher. I've got a custom loop on my test bench, and I'm pulling 5.0GHz+ overclocks on Sandy Bridge CPUs, 4.7GHz+ on Sandy Bridge-E, and 4.7GHz on Ivy Bridge. The latter of course has thermal issues once you crank the voltages up.

Modifié par Dead_Meat357, 18 mai 2012 - 04:49 .

[InBleedingRapture]

Wow. Yes, I actually did have a decent setup (Swifteck pump with 3/4" barbs, a DangerDen block, and two rads), but I think it was my tubing that killed it. At 3/4" it's hard to find tubing that doesn't kink. I tried coiling wire around it, etc., to no avail.

If you want some spare water cooling parts, I'd be happy to send them to you, free of charge. The DD block was made for an LGA775 socket, but I modded it to work with my i7. Just a matter of drilling new mounting holes, as the copper base is exactly the same size.

Let me know via PM if you want a few spare parts.

[Dead_Meat357]

InBleedingRapture wrote...

Wow. Yes, I actually did have a decent setup (Swifteck pump with 3/4" barbs, a DangerDen block, and two rads), but I think it was my tubing that killed it. At 3/4" it's hard to find tubing that doesn't kink. I tried coiling wire around it, etc., to no avail.

If you want some spare water cooling parts, I'd be happy to send them to you, free of charge. The DD block was made for an LGA775 socket, but I modded it to work with my i7. Just a matter of drilling new mounting holes, as the copper base is exactly the same size.

Let me know via PM if you want a few spare parts.

I've got more water cooling hardware than I know what to do with as it is.

[InBleedingRapture]

Haha, no prob. Maybe I'll take another stab at it one of these days. My new tower sits at about 2-4 degrees above ambient, so perhaps I'll have better luck.

[neilthecellist]

I'll take it. Send me an email at neilthecellist@gmail.com. Free shipping or charge?

[Rahabzu]

My head exploded.

[ZackG312]

SliPaladin wrote...

So basically at 1080p my 1gig GTX560Ti wont be enough... damn :/

I have a i7 2600 8gb ram and GTX 560 ti 1gb. and have all files loaded at once and the game hasnt shut down on me.

[Desuke]

ZackG312 wrote...

SliPaladin wrote...

So basically at 1080p my 1gig GTX560Ti wont be enough... damn :/

I have a i7 2600 8gb ram and GTX 560 ti 1gb. and have all files loaded at once and the game hasnt shut down on me.

Its all depends on the main memory system you have I guess. At least 8 gig.

[Desuke]

Its all depends on the main memory system you have I guess. At least 8gig.

[neilthecellist]

Got your message, Rapture!

EDIT:

Desuke wrote...
Its all depends on the main memory system you have I guess. At least 8gig.

Not necessarily, but having 8 gigabytes of RAM is pretty much a minimum standard for gaming on a x64 machine that isn't Windows XP. In my experience.

Rahabzu wrote...

My head exploded.

That's a shame, considering you didn't have to compute anything or discuss anything like a literary scholar or an experienced tech. Anything positive to contribute?

Modifié par neilthecellist, 18 mai 2012 - 11:30 .

[muchevo]

hi smarteck
thanks for this awsome work
BTW can i repost it to another forums?

[SliPaladin]

ZackG312 wrote...

SliPaladin wrote...

So basically at 1080p my 1gig GTX560Ti wont be enough... damn :/

I have a i7 2600 8gb ram and GTX 560 ti 1gb. and have all files loaded at once and the game hasnt shut down on me.

Good to know, my specs are exactly the same. 2600k overclocked to 4.5Ghz, 8GB 1600Mhz ram

Modifié par SliPaladin, 18 mai 2012 - 12:19 .

[SliPaladin]

Where do I put the batch file? I'm kinda noobish with these things. Also you use notepad for the text file, right?

[neilthecellist]

Binaries/Win32

[Fredy AG]

muchevo wrote...

hi smarteck
thanks for this awsome work
BTW can i repost it to another forums?

Of Course. 

Where is it?

//////