X79 and/or rampage IV OC'ing thread for those of us still left....

[Kanan]

Aww dont be so negative (or old) man, you won't lose data, and Win 10 is better, it's no "risk". It's maybe a "risk" if you think it is, and mess it up because of it. Same as with fear, if you have fear falling off the bike it will happen evantually. Plus there are 10.000 ways to backup and rescue things. If you want to stick to Win 7 do it, no prob, just your arguments are so lame, it's out of reality. Maybe it's lame from me to call you a Granny, I think Grannys don't behave like that. hahahaha! Okay okay, I'll stop it.

 

[dalekdukesboy]

Hey, all of us have lost data, had programs crash, hard drives fail, gpu's/cpu's either blow up or have issues. Truthfully, yes you can backup etc, but somewhere you missed I kinda had to replace everything in this system minus the ES, everything else is new (to me). So last thing I care to do or my wallet cares to do is get another operating system I don't need, whatsoever. ALso I've taken a lot of risks or change with this system, including trying to change a fan while it was running and fried a card. The risk is minimal, simply mentioned it, been under the hood so often with this pc literally and figuratively with settings honestly I'm kinda just happy it works atm. So I played a game tonight and actually enjoyed the fruits of my labor.

 

[dalekdukesboy]

This is my best result thus far. Actually had memory at only 2005 and core was bit higher than last one but memory 50 mhz lower and results increased showing just 10 mhz or so increase on the core makes a much larger difference than memory least on this benchmark...though that is fairly universal in most things core makes biggest difference memory has less effect per mhz increased.

I love how in benchmarks and any software that tries to read my cpu it shows up as unknown and totally flummoxes the program because it's an ES. I just find it humorous.

 

[Kanan]

It's funny yeah because Intel wrote the wrong CPU code/name in it. Normally I think most ES have same code as retail model just with "(ES)" behind it.

 

[dalekdukesboy]

Either way it's humorous, and long as it works I'm good with it.

http://www.xtremesystems.org/forums...ent-IC)-G-SkILL-ASUS-Rampage-IV-Extreme/page9

This is a good article if/when you touch the memory on your pc

 

[Kanan]

Either way it's humorous, and long as it works I'm good with it.

http://www.xtremesystems.org/forums...ent-IC)-G-SkILL-ASUS-Rampage-IV-Extreme/page9

This is a good article if/when you touch the memory on your pc

Yeah thanks I'll come back to it if/when I want to OC my memory. So far it doesn't seem so, but who knows how long I'll use this system and how much bandwidth is needed in the future.

PS. Here's my 3DM11 run:
http://www.3dmark.com/3dm11/11414355
With i7 3820 at 4.3GHz:
http://www.3dmark.com/3dm11/11407318
Comparison:
http://www.3dmark.com/compare/3dm11/11414355/3dm11/11407318#

Firestrike:
http://www.3dmark.com/fs/9331757
With i7 3820:
http://www.3dmark.com/fs/9287606
Comparison:
http://www.3dmark.com/compare/fs/9331757/fs/9287606

Maybe interesting.

 

[dalekdukesboy]

Benchmark obviously likes the couple extra cores, not bad at any rate. Those were pretty good cards when they were made and obviously at least going by that score shows it probably can run almost any game at reasonable settings.

 

[dalekdukesboy]

I unhooked fan I had loosely placed over VRM, and quickly found out by touching it how much hotter it runs without it. Also can't get past 1 run of burntest where before I could pass ten so an fyi this platform even with a high end mobo like mine really needs a fan over that area of motherboard if you do anything for memory overclocking or high mhz cpu core clocking as well I'm sure it would get hot.

 

[Kanan]

Well maybe so but I think because you have bad airflow there. Watercooling doesn't help with that (the opposite really) and I'm not sure if your case is good either - I just don't know it. I know that I have a lot of airflow. I have 2x140mm blowing on hardware directly and 1x 140mm sucking out besides the NH-D14 that helps too with component cooling.

Essentially you're just "repairing" the damage you have done by removing the proper air cooler and replacing it with aio water. These boards aren't made with that in mind, high overclocks and then no component cooling? Nah I guess you're supposed to cool it yourself then. And because you do it its working well - until of course you stop it.

 

[dalekdukesboy]

Yeah, kinda silly when you think of how much this board cost new and that it was sold and advertised as a serious overclocking board, THE overclocking board for x79 really. Case has good airflow but not for that area particularly since I removed couple fans on the back and the fans there now are on the RAD so they are not sucking much from case they are more just blasting through the RAD and due to the cases' size it actually is a minus because there are dead spots because fans can't hit every spot. The case has good fans/flow but again the size of it works against it, plus wires/hard drives/other obstructions block a good portion of the fans so it has the potential and has 7 fans on it stock...but the VRM on my board is near top of the case and back/middle...placing it many inches below big 220 mm fan on top of case and set back far enough on bottom of case (obviously) but fans on case are further to side of case so they don't really hit that area well at all. I now have a serious fan tie-wrapped in place right over it where before it was askew a bit and not hitting as directly. Yes, the watercooling only exacerbated the problem for I lost 2 fans or 3 at times from the nh-d14 which were over that area, now just a waterblock. However now putting that fan there it's way better than it was even with the air cooling which was only a passive amount of air. Now it basically has a hurricane on the vrm lol.

 

[dalekdukesboy]

OH, separate note...that msi gaming cooler? I just checked it out when I pulled out my card and put fan back on vrm etc etc, the fan connectors on it are not standard, one fan has a weird 5 pin connector the other one has 3 I think. The typical connection I see on these is 4 pin or whatever was on the armor cooler I took off and now the accellero cooler I put on, both had same connection. So seems MSI was sneaky with that so you couldn't just take their cooler OR their gaming PCB and mix and match different PCB's or coolers on them. I considered trying the gaming cooler and that is when I noticed the connections I didn't really pay attention to when I looked at it before.

I reset the arctic cooler and used good tim on it so hopefully it's a bit better than it was, before I just slapped it on with old tim and was seeing if it worked lol.

 

[Kanan]

I don't think it has to do with sneaky practices. It's another cooler with passive on/off mode, maybe that's why it's different - I don't think Msi gives a shit about 0.1% users that swap a Msi cooler with another Msi cooler from another Msi card.

So how much of a difference did the Arctic cooler make compared to the armor cooler?

 

[dalekdukesboy]

I don't think it has to do with sneaky practices. It's another cooler with passive on/off mode, maybe that's why it's different - I don't think Msi gives a shit about 0.1% users that swap a Msi cooler with another Msi cooler from another Msi card.

So how much of a difference did the Arctic cooler make compared to the armor cooler?

Perhaps, but trust me it's a weird couple of connections so granted I as well wondered why even do it but that's only reason other than mechanically they had to, but I'm not convinced, also I wasn't thinking MSI to MSI I was thinking non-MSI obviously. Anyway that was just what I found out and means that cooler is useless without adapters if I could find them or another MSI card with same connections. The armor cooler was barely adequate....that is why I changed it. At first I thought it was ok but over time in 3d it would get into the 70's even when you cranked the fans. It's a neat little compact cooler, BUT not for a 980 ti more for a 980 or less with much less wattage to diffuse. I just tried the arctic cooler after re-applying it with Noctua TIM (obviously had some from my old cooler) and it was already better than the armor cooler but now it's plain amazing. I'm very glad I reapplied the tim and made sure it made good contact and cranked the screws down. I ran a minute or two of furmark....now the card even overclocked and volted to 1.224 volts was running in the 40 degree Celsius range lol. I'll do a longer run just to see how cool it stays over time, but way better than before. Needless to say heat won't be an issue on that card anymore, might get ice on it now. Also just proved my theory that the vrm was and is THE key to overclocking stability on this board. Just ran 2400 mhz at 10-11-11 timings totally stable.

 

[dalekdukesboy]

Now I just have to get sandra or some other benchmark for memory bandwidth etc lol. Now I may with my memory and 4.6 ghz on cpu and 16 gigs in quad channel might get some respectable results.

 

[Ithanul]

I will post my X79 when I get it into a case and under water. I just being a lazy butt lately. Then again, been busy with class homework. Had some serious trig questions I had to do (proving identities). Talk about taking a good 10+ mins depending on the equation.

On the whole W7/W10. I'm just waiting it out. I tend to only move when it necessary to do so. Plus, at the rate I even game, I can take my time. I just finally a few weeks back got around to playing Bioshock Infinite. Heck, I still need to play the first Tomb Raider. Yes, I'm slow. I blame it on my 3DS and Wii U. Usually what makes me move is software for digital drawing. But so far Art Rage, Photoshop, etc. still work fine on W7.

 

[dalekdukesboy]

Thanks, also based on basic info from both of you about school and general conversation as an aside I'm guessing I probably am considerably older. Trig, ick, I stink with math...much more adept with English and writing. Anyway thanks for the input! I am actually downloading a couple benchmark suites I can play around with like this...[/ATTACH]

Pretty impressive results, not sure the timings etc of the ddr3 and ddr4 results on their chart but my memory at current stable settings crushes all their baselines. Admittedly wasn't thinking I'd be ahead of any ddr4 especially with latency and such.

 

[Kanan]

Well I moved before the free period of win10 expired that's it - sole reason more or less. I even waited extra for it, moved in the last days of it. Only new and really usable thing is maybe time spy bench for now, apps are better too but I rarely use them. Then again I had win 8.1 before it - win7 lol kinda oldschool os, simply lacks a lot and is old. Win 7 users don't accept that but it's a fact they downplay it. A neutral user who knows everything will see it differently though.

@dalekdukesboy you really need custom water cooling for a user like you its the only suitable thing on the long run.

BTW latency is the only thing that sets ddr3 apart from ddr4 - that and voltage that's lower on ddr4. I'm kinda wondering how you are surprised your ddr3 timings are better than ddr4.
Always the same. every new generation of ram buys more bandwidth by trading in latency, that's why I don't overclock I stay on fast timings, it's not really different to 2400 + slow timings. Plus I don't need more bandwidth that's the only thing I don't need with a quad channel system.

 

[dalekdukesboy]

No I figured the ddr4 was fairly high clocked....it was 2133 mhz lol no wonder it didn't show well. I hate it when they put baselines that are irrelevant like that. So yeah I'll have to check and see if there are better results to compare to with that program. What I really want to see is how I can compare to high mhz clocked memory ddr3 and ddr4. So I was only surprised because I expected ddr4 at a decent clock not 2133 which is crap speed for ddr4 especially if the timings are bad as well. So yes latency in particular is where I figured it would have results that would trump mine, but of course they didn't there weren't any comparable memory results.

 

[Kanan]

Well the magic is to get high clocks with fast latencies, if you can do that, it's a pure win. 2400 + CL10 for example, that's really good. DDR4 3000 has a very high clock but it's latencies are slow, it's simply made up by the higher bandwidth. I think Anandtech sometime made a test, latency vs bandwidth and whats essentially the sweetspot. For DDR3 it was 1866/2133/2400 + fast timings (1866 = CL9 or even 8, 2133/2400 CL10/11), DDR4 is only good above 2800 or 3000. I don't see the point in buying 2133 DDR4 or even 2400 DDR4, it's really pointless. Buy 2800 or 3000 minimum or go home, essentially.

This is what I mean by "3000 or go home":
http://www.anandtech.com/show/8959/...-3200-with-gskill-corsair-adata-and-crucial/8
There are other tests too.

 

[dalekdukesboy]

Well the magic is to get high clocks with fast latencies, if you can do that, it's a pure win. 2400 + CL10 for example, that's really good. DDR4 3000 has a very high clock but it's latencies are slow, it's simply made up by the higher bandwidth. I think Anandtech sometime made a test, latency vs bandwidth and whats essentially the sweetspot. For DDR3 it was 1866/2133/2400 + fast timings (1866 = CL9 or even 8, 2133/2400 CL10/11), DDR4 is only good above 2800 or 3000. I don't see the point in buying 2133 DDR4 or even 2400 DDR4, it's really pointless. Buy 2800 or 3000 minimum or go home, essentially.

This is what I mean by "3000 or go home":
http://www.anandtech.com/show/8959/...-3200-with-gskill-corsair-adata-and-crucial/8
There are other tests too.

Yeah that article is ok but maybe on quick perusal I only notice them comparing the ddr4/ddr3 with same exact timings/speed etc for apples to apples comparison...rest of article was about ddr4 only. Either way I have seen a few articles where they have put them head to head oc'ing them like I did and then the same with ddr4 and came up with results. Thanks by the way, yes I'm pretty psyched I got this stuff to do 2400 with ease and now with good timings! I was trying before and with the air cooling even that spot of the vrm was always quite hot and I noticed throttling when I pushed it or failures/blue screens etc and just much inconsistency even from one try to the next. I can safely say the board was the weak spot and there is only a passive heatsink on that area and it just doesn't do enough on its' own without a fan to deal with overclocking, period. Thankfully I ordered 3 of the fans even once I realized I only needed 2 for the watercooler because they are just great high cfm fans with relatively low noise vs cfm output so I knew I'd use it somewhere!

 

[dalekdukesboy]

Soon our lazy butt lady friend will have competing numbers for us I'm sure with her watercooling (way cool) and I'll have competition see what we can do. Yeah technically with ivy bridge you could get up to 3000 mhz believe it or not with ddr3, least you could buy modules that fast of course it depended on your CPU memory ic and what it would do but I obviously have a good memory controller on this cpu. I'm pleasantly surprised the weak link was NOT my cpu and simply the board circuitry (which I did run into before on a nvidia sli board way back when).

Wow I just purchased aida 64 since that has lots of good benchmarks etc including memory...I will figure out how to get a more concise printout of results but for now let's just say results are quite good at present settings.

 

[Kanan]

Sounds good, I'd like to see some pictures of that!

Yeah there were better tests I didn't find the link, maybe later if I have the time.

Yep, DDR3 can hit up to or over 3000. DDR4 really is more about long term higher capabilities and less energy consumption (that's obviously not really important to normal users, but for servers it is).

 

[dalekdukesboy]

Well if you meant pictures of the Aida results I have them, you can print reports but even longer than just taking individual screenshots I could just type in results but you need to compare it to other systems which here you can see that...excellent results all around...actually fantastic, except for my memory write times are fairly weak not sure why could be more dependent on timings than speed? Only reason I could see an x79 system clocked lower with slower memory but tighter timings being better.

ok NM the write is there...looks like I'm missing the read...

 

[dalekdukesboy]

There's the read and the cache and memory benchmark. One thing that is holding it back I think is I'm on 2 timings, obviously If I have 1n timings it would boost results just not sure if I can get it to work at this speed, but I should give it a go to see.

*Just edited* put the 2t vs 1t in that order one over the other all other settings/timings were the same, not a massive increase but does ad several thousand to each memory score.

 

[dalekdukesboy]

no idea if it's stable but 1n timings is at the least benchmark stable...still can't figure out why the write speeds are not so good...not sure if that's just a glitch or what. Just doesn't seem to be in line with other results and what I'd expect...anyway here are some 1n timing results, not huge improvements but there are a few jumps to very top of charts with it...