Battlefield 4 loves high speed memory

[SPBHM]

when Corsair blog is your only source I'm a little skeptical, but it might well be real I guess, but it could be limited to some very specific conditions, and 780 SLI + 4770 4.4GHz is pretty fast and can give a false impression, if you have a single card, a slower CPU, using your money for faster memory might not be the most efficient route,

anyway, BF4 don't need more than 8GB, so going for more speed than capacity is probably a good thing here.... but... more testing is needed.

 

[Majcric]

Not to mention more games need to utilize higher memory freq or to me it's just not worth it for one game. Nonetheless, interesting results.

 

[Slomo4shO]

Enthusiasm is not enough to get hw to 4.8 for more than 5%
Most wont go past 4.3 and some wont even go there.
Is it right to delid your cpu and do personal binning to get the effect of ram in bf4? Not real testing imho. Unserious.

https://docs.google.com/a/overclock...NjlDYWl6ZnV4OVlNc0lMU1V3c1E&usp=sharing#gid=0

I didn't delid my chip, and I see more than enough 4.4GHz or higher chips listed on an enthusiast website to believe the 4.4GHz used in the test is more than fine.

Paul from newegg did an auto overclock run with I think 6 or 7 i7-4770ks using an ASUS board and all of them did more than 4.4GHz with less than 1.3v a few were hitting 4.7/4.8.

If you want to perpetuate a myth by all means, I only offer contradictory evidence to your words.

Hmm what to believe... random overclocking claims without any proof, a volunteered submissions of results from a sample of 88 that lacks random sampling, or Asus's sampling of hundreds of Haswell chips which concluded that

70% hit 4.5GHz, 30% reached 4.6GHz, and 20% made 4.7GHz. Only 10% were stable at 4.8GHz

The above results imply that chips that clock at 4.4GHz or lower only make up 30% of the total chips. Since when was 30% a majority?

 

[BallaTheFeared]

Sorry I couldn't find Paul's video, I would have submitted it had I been able to locate it. I don't know his personal channel and his newegg channel is all I can get hits for.

I refined by search on youtube and was able to locate the video.

http://www.youtube.com/watch?v=WugDJbtBwwQ

He tests core with 1600 memory, 2400 memory and 2666 mem.


ASUS comments are based on pre launch samples, though still entirely valid their sample size and scope at the time is severely limited as it stands today.

 

[HurleyBird]

This should be in the CPU forums, no?

 

[BallaTheFeared]

This should be in the CPU forums, no?

This thread transcends those limitations.

Is it CPU? Is it GPU? Is it RAM?


/mindblown

 

[Carfax83]

This should be in the CPU forums, no?

I posted the same thread in the memory and storage forum, which is where it belongs.

 

[Slomo4shO]

ASUS comments are based on pre launch samples, though still entirely valid their sample size and scope at the time is severely limited as it stands today.

Unless there is supporting evidence that Intel's production techniques have improved or changed since the pre-launch, a random sample size is still able to provide an accurate depiction of realistic overclock potentials of the complete population of Haswell CPUs based on a limited sample size.

However, sampling from the Overclock.net forums is not a true reprentation of the market and, therefore, can't accurately predict the true overclocking bell curve because individuals who perceive themselves to have bad or great chips are more likely to volunteer to submit their results. In addition, these results have high variance in cooling, hardware, and technical knowledge so the the lack of control in variables also takes away from the validity of any such pools of data.

Interestingly enough, if you confine the overclock results to 1.3V, you notice that only 3 samples out of the 46 hit 4.8GHz or higher and this number goes to 6 out of 66 when observing the results with the voltage limited to 1.35V

 

[BallaTheFeared]

Well I guess we'll disagree, since places like ocn, here, and any other enthusiast forum is where the rubber meets the road. Which is what the other guy was saying 4.4GHz was unrealistic for haswell for end users, the high variance of actual user configurations and willingness to use more voltage results are exactly what i posted.

Either way why are we even arguing over it, point still stands 4.4GHz is a valid and perfectly reasonable clock speed for Corsair to have used.

 

[skipsneeky2]

My 2500 non k has felt pretty mighty till BF4 came out,the 1333 memory and the lack of overclocking or ht of the i7 really makes this thing look pale in comparison to something like a overclocked 4770k.

God i hope the final game doesn't murder this chip.

 

[Slomo4shO]

Either way why are we even arguing over it, point still stands 4.4GHz is a valid and perfectly reasonable clock speed for Corsair to have used.

That wasn't the argument I was pointing out that neither of your two conclusions were based on accurate data. 4.5GHz appears to be the the mode and is also a realistic clock that a majority of the population can reach with sufficient cooling so there is no reason to argue against the validity of using 4.4GHz overclock for benchmarks. However, there are plenty of other issues with the review that bring forth questions about the results. The blog appears to be a marketing ploy more than anything and I would personally like to see someone run these tests that didn't have anything to gain from such findings.

 

[BallaTheFeared]

In your opinion, personally I value user data be it cpu or gpu or any other overclockable hardware.

Why you put so much stock in what ASUS got from 500 samples pre release is your concern, not mine and it in no way changes what I think because I was aware of their data before I even posted I just didn't value it the way you do.

You are "suspecious" of these RAM results because they come from a RAM company, but you put all your chips on the overclocking results of a motherboard company that makes bank selling overclocking motherboards. If you can't see the problem in your own train of thought, again that is your concern not mine.

 

[Slomo4shO]

I just put more stock in statistical data that can be reproduced and is unbiased. The Asus sampling isn't perfect by any means but it is the best data set available currently. User data is great if it is obtained through random sampling but is statically worthless when it isn't. Math doesn't lie, the data collected might.

 

[Majcric]

My 2500 non k has felt pretty mighty till BF4 came out,the 1333 memory and the lack of overclocking or ht of the i7 really makes this thing look pale in comparison to something like a overclocked 4770k.

God i hope the final game doesn't murder this chip.

Yeah I know what ya mean. I need to get to work on my chip in the near future.

 

[BallaTheFeared]

I just put more stock in statistical data that can be reproduced and is unbiased. The Asus sampling isn't perfect by any means but it is the best data set available currently. User data is great if it is obtained through random sampling but is statically worthless when it isn't. Math doesn't lie, the data collected might.

This is all I'm going to say after this, we're derailing at this point.

How did you, me, or anyone else know the 7950 was a good overclocking card on average?

Neither ASUS nor anyone else released anything similar to what you're quoting, yet all of us will agree many of AMDs chips were very good for overclocking this generation.

So how did we come to that conclusion given our data is purely user based with by comparison a trivial amount of reviewer sampling and no amount of large quantity sampling by a single source?

 

[Arkaign]

Yeah I know what ya mean. I need to get to work on my chip in the near future.

You're running stock though, I'd be honestly stunned if you couldn't get an IBT-stable 4.5Ghz after popping a $30 212 on there, at low volts to boot. I've literally never seen a SB that couldn't do 4.5, at which point you're still in perfectly good shape to pair with a single 680.

 

[Slomo4shO]

How did you, me, or anyone else know the 7950 was a good overclocking card on average?

By using the best set of data available.

It also doesn't hurt that just about every review was able to overclock the 7950 by 28-35% above the 800MHz reference clock at stock voltage:

http://www.techpowerup.com/reviews/AMD/HD_7950/30.html
http://www.techpowerup.com/reviews/Sapphire/HD_7950_Vapor-X/31.html
http://www.techpowerup.com/reviews/HIS/HD_7950_X2_Boost/31.html
http://www.legitreviews.com/sapphire-radeon-hd-7950-3gb-vapor-x-video-card-review_2050/13
http://www.overclockersclub.com/reviews/sapphire_7950_vaporx/4.htm
http://www.guru3d.com/articles_pages/msi_radeon_hd_7950_twin_frozer_iii_review,24.html
http://www.overclockersclub.com/reviews/xfx_sapphire_hd7950_3gb/6.htm
http://www.hardwarecanucks.com/foru...2-sapphire-hd-7950-dual-fan-oc-review-13.html
http://www.overclock3d.net/reviews/gpu_displays/sapphire_hd7950_overclock_edition_review/3
http://www.guru3d.com/articles_pages/powercolor_radeon_hd_7950_pcs_review,24.html
http://www.techpowerup.com/reviews/AMD/HD_7950/30.html
http://www.tomshardware.com/reviews/radeon-hd-7950-review-benchmark,3207-13.html
http://www.anandtech.com/show/5476/amd-radeon-7950-review/17

 

[skipsneeky2]

Yeah I know what ya mean. I need to get to work on my chip in the near future.

Oh i had the cash for a i7 3770k and at one point i already had a maximus gene z77 mobo in my rig and i just needed to purchase the 3770k,i went out of town and some short circuit in my wall took out the board and a friend of mine tried to fix it and bent half the pins by mistake.Came home to find my 7850 was half dead as well and eventually the cash for the i7 went towards the mobo and ram in my signature and a temporary gtx650 as basically the board,memory and the 7850 went out.:|Even got a i3 3220 as i feared my i5 was finished but luckily not.:awe:

 

[omeds]

Enthusiasm is not enough to get hw to 4.8 for more than 5%
Most wont go past 4.3 and some wont even go there.
Is it right to delid your cpu and do personal binning to get the effect of ram in bf4? Not real testing imho. Unserious.

https://docs.google.com/a/overclock...NjlDYWl6ZnV4OVlNc0lMU1V3c1E&usp=sharing#gid=0

I didn't delid my chip, and I see more than enough 4.4GHz or higher chips listed on an enthusiast website to believe the 4.4GHz used in the test is more than fine.

Paul from newegg did an auto overclock run with I think 6 or 7 i7-4770ks using an ASUS board and all of them did more than 4.4GHz with less than 1.3v a few were hitting 4.7/4.8.

If you want to perpetuate a myth by all means, I only offer contradictory evidence to your words.

FWIW I have a delidded 4770k at 4.8 too, with 2666MHz RAM. First chip too, did not bin.

 

[lavaheadache]

This thread transcends those limitations.

Is it CPU? Is it GPU? Is it RAM?


/mindblown

PC Gaming Sub forum....