Disappointed with performance

[surgetheurge]

Hey.
So a year ago, I made a PC for a youtuber, (whom actually didn't pay me for it, so they scammed me for 3000 dollars, but thats not the focus here) and this pc used the 4930k processor with the P9X79 WS motherboard. For some reason i thought I needed a ws motherboard... Don't remember why...

Anyway I cranked it up to 4.8 Ghz and water cooled it with h100i.
I took the http://www.passmark.com/ software, and tested the processor. It hit 16,000 points on 4.8 ghz. i think 15k something on 4.6. 4.8 ran wayyyy to hot. Like everything was hitting 86c so it set it to 4.6 where it never hit 80.

So anyway...my current cpu was the 8350 hitting 9.1k points @ 4.2 ghz. Knowing the 4930k hits 16k points. I decided to spend a thousand dollars on upgrades to get that sweet 16k points, so I can render videos faster.

I got the sabertooth x79 motherboard, with the 4930k processor, and h100i. Cranked it to 4.6 ghz. I think it can't do 4.8. Might have tested 4.7 too. Anyway ran the test. got 12k points. I wish you could see my expression lol.

I currently use cheap 60 dollar stick gskill ram. Could the ram be a factor? Also I am using the x79 sabertooth motherboard, previously, the p9x79, it uses some-kind of a express chipset that boosts sli performance by 30%. I called up asus and asked them if that chipset could boost cpu performance and they said no. Could the chipset be a factor? just in case the support people don't know better x.x

So yea, my previous computer had a 60 dollar stick of ram, 60 dollar motherboard, 40 dollar heat sink, and a 190 dollar cpu. This computer hit 9k points on passmark. My new computer has the 780ti, 4930k processor, 200 dollar motherboard, and a 500 dollar scorpion deluxe SSD (1300 MB read, 500 MB write, kind of pissed about this, advertised to be 2000/2000, CrystalDiskMark says its actually 1300/500)

30% more performance isn't worth the thousands of dollars of upgrades lol.
Why can't i get the same numbers with this cpu as I did with the previous one?

 

[Dicehunter]

It can be multiple things, Newer/older software version, Programs running in the background, Out of date drivers.

Is the passmark benchmark the exact same number version ?

Also do you have Intel speedstep and the energy saving settings disabled to maintain the clocks etc... ?

Did you run just the CPU benchmark or others as well which would in fact lower the score ?

 

[surgetheurge]

It can be multiple things, Newer/older software version, Programs running in the background, Out of date drivers.

nothing running in the background

drivers for motherboard? should i update my drivers?

Is the passmark benchmark the exact same number version ?
[QUOTE/]
I don't remember the passmark version.

Also do you have Intel speedstep and the energy saving settings disabled to maintain the clocks etc... ?

yea



Still all this stuff seems minor... i think it might even have been 18k at 4.8 and 16k at at 4.6. Were talking about a massive drop in performance, surely this must be caused by some kind of bottleneck somewhere. I can't believe drivers would cause like a 40% performance drop. ... but I really probably don't know better.

 

[remember300]

most things are not as good as they say on the tin, unfortunately, have you tried a dif sata 3 port?
Could be a Bus delay or your ssd.
plus do i have to mention silicon loto.
Tried re seating you cooler?
are your fans on full are you also at the same room temp?

 

[tinytomlogan]

Theres loads to set up and tweak with 2011 chips

youve not mentioned vcore once and thats what causes heat - PLEASE tell me it wasnt on 'auto'

 

[surgetheurge]

Theres loads to set up and tweak with 2011 chips

youve not mentioned vcore once and thats what causes heat - PLEASE tell me it wasnt on 'auto'

hehe. it was xD
Sorry for eating up my time with my noobishness... I am going to go look up an overclocking tutorial on the x79 sabertooth motherboard.

 

[MacLeod]

Isn't Passmark a synthetic benchmark as in it doesn't really mean anything kinda like 3DMark?

Synthetic benchmarks like that shouldn't really be relied on all that much for judging how well your rig performs. My 8350 get a 7.8 on the Windows Experience thingy. That's 1 tenth of a point from a perfect score on a 2 year old (albeit heavily overclocked) CPU. It doesn't mean much.

Start using the applications you were using with your old rig and see if you see a bigger difference. I'm betting you will.

 

[surgetheurge]

Isn't Passmark a synthetic benchmark as in it doesn't really mean anything kinda like 3DMark?

Synthetic benchmarks like that shouldn't really be relied on all that much for judging how well your rig performs. My 8350 get a 7.8 on the Windows Experience thingy. That's 1 tenth of a point from a perfect score on a 2 year old (albeit heavily overclocked) CPU. It doesn't mean much.

Start using the applications you were using with your old rig and see if you see a bigger difference. I'm betting you will.

I am trying to minimize rendering time of videos. Most video editing programs don't use gpu acceleration, the video editing programs that use gpu acceleration, only use like 1% load.
A synthetic benchmark loads every core, every thread to 100% cpu usage right?
usually when rendering videos the same thing happens, 100% everywhere. So the process of rendering videos works similarly to synthetic benchmarks, therefore synthetic benchmarks can be a very good way to measure rendering speed.

feel free to correct me if i am wrong. I am no expert.

 

[MacLeod]

Looking at benchmarks on various sites, it looks like 30% is about ballpark the difference you're gonna get between a 4930K and 8350 so what you're getting may be about right.

The thing with synthetic benchmarks like 3DMark is that they don't always translate to real world differences. In the first 8350 vs i5 reviews, you had the i5 beating the 8350 by double in some benches like 3DMark but when actually gaming, the differences were much much smaller and when review sites started using real world benches and not benching at 800x600 resolution, the differences turned out to be pretty much nothing.

For video rendering for example the 8350 is actually a pretty competent chip and that is right in its wheelhouse. Intel has a huge advantage in performance per clock single threaded but when you use stuff that can use all 8 cores of a FX especially when overclocked, it's a pretty solid performer so you may be expecting a little more performance than you're actually gonna get. Don't get me wrong, a 4930K is a WAY better CPU than a 8350 and a 30% bump is pretty huge. You gotta think, each Intel generation has only brought around a 5% performance increase so wanting a 50-100% increase from a 8350 to a 4930K might be expecting a little much.

 

[surgetheurge]

Looking at benchmarks on various sites, it looks like 30% is about ballpark the difference you're gonna get between a 4930K and 8350 so what you're getting may be about right.

The thing with synthetic benchmarks like 3DMark is that they don't always translate to real world differences. In the first 8350 vs i5 reviews, you had the i5 beating the 8350 by double in some benches like 3DMark but when actually gaming, the differences were much much smaller and when review sites started using real world benches and not benching at 800x600 resolution, the differences turned out to be pretty much nothing.

For video rendering for example the 8350 is actually a pretty competent chip and that is right in its wheelhouse. Intel has a huge advantage in performance per clock single threaded but when you use stuff that can use all 8 cores of a FX especially when overclocked, it's a pretty solid performer so you may be expecting a little more performance than you're actually gonna get. Don't get me wrong, a 4930K is a WAY better CPU than a 8350 and a 30% bump is pretty huge. You gotta think, each Intel generation has only brought around a 5% performance increase so wanting a 50-100% increase from a 8350 to a 4930K might be expecting a little much.

Well how the hell did the 4930k hit 18k points at 4.8 ghz? O.O

 

[surgetheurge]

another important thing worth noting:
Every time i start up the computer, it says overclocking failed press F1. Then I go into bios, set it to the same exact settings, and when computer starts up, it has the overclock and it runs stable. If I restart it, same thing again, overclocking failed press F1.

I got so bothered with it, I set the settings to default.

Also, I tried the AI Tweeking tool to overclock the computer for me, and I also get the same press F1 message.

edit: whoops. Sorry for the double post.

 

[MacLeod]

Have you stress tested this overclock with Prime95 or OCCT? It could be that your overclock isn't stable and there is some throttling going on or something like that.