• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

new member question

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

R_Pierce

Member
Joined
Feb 4, 2017
Location
Marion, IA
Hey guys, just started finally messing around on HWbot. I just entered the Rookie Rumble #43. Im currently sitting in 11th place overall. I think I can muster up some better numbers, just not exactly sure what direction to take.

Linked below is the competition:

http://oc-esports.io/#!/round/rookie_rumble_43

And here is an example:

image_id_1820532.jpeg

Same processor, but running at an effective 5.2 (51 multiplier, 102 BCLK) vs my 5.1. He is running his cache at 5.0 and mine is currently at 4.8 (its stable at 4.9 as well, but have not went above that) But the difference in score is almost 200 marks. I noticed he does have faster DRam. But also he is running his at 1T (2000 12-12-12-26-1T) Mine is running at 1800 15-15-15-36-2T. Would this make that big of a difference? I havent really been able to get my RAM to OC much past what the XMP values are. I can get it to boot at 4000 21-21-21-36-2T with a 1.4000 dram voltage, 1.300 VCCIO 1.300 SA and current draw set to 130%. But it isnt stable even enough to finish the XTU bench run.

Thanks in advance for any help!!

Rodney
 
Last edited:
I also thought about going down to 2 DIMMS instead of 4, just to see if it helped with access times or clock speeds (assuming this would help my XTU bench?)
 
Reducing the number of sticks will always improve overclockability. The IMC(integrated memory controller) is your biggest enemy when pushing a lot of RAM. If you genuinely want to push your rank pull the extra sticks bump the volts and tighten the timings. 1T VS 2T is a fair improvement, but getting generally lower timings and a higher MHZ on the RAM will generally yeild better results vs just going from 2T to 1T timing.

As for the memory tmings 2000Mhz @ 12-12-12-16-1T is substantially faster than 1800Mhz @ 15-15-15-36 2T. The thing with competitive benchmarking is you have to accept that if you want to win you may have to sacrifice some hardware in the process. Given that this is a short benchmark you will fair better if you throw caution to the wind and dial it up to a 12. Whatever you think a fair or save voltage limit is for any piece of hardware add at least 0.2V onto that. Pull the sides off your case and point the biggest fan you have into the mb chamber. World records are made by hardware that is so close to the edge of crashing that you may have to run a bench multiple times just to get it to last long enough for a valid screen shot. Hes crushing you at 5.1 shoot for 5.2 minimum. He has 200Mhz on your RAM speed and has tighter timings push those ICs to their limit and crank it to the same or better speed, and give them whatever voltage your MB will let you.

In this world its not about stability its about getting the screen shot. Even if it crashes 0.1 seconds later you still got your run done. Now this practice is not ok for long term overclocks, but this is a competition after all. The few minutes it takes to complete the boot, bench, and save results process is unlikely to kill your hardware. However, that does not mean that its safe. You always risk killing something when overclocking, but for the sake of glory and the very low likelyhood that it will go down like that we are usually willing to push it.
 
Copy that! I'm on it when I get off work in the morning. Also, what are you guys using for a quick screen capture utility? Something that I don't have to worry about using the snipping tool and taking that time.

 
If you can run 4 memory sticks at about the same clock and timings but CR2 then it will be probably faster than 2 sticks at CR1. Still in XTU you need high memory clock and tight timings to compete with others. Most users run it on samsung b 2x8GB ~3866-4000 12-12-12 ( voltages above 1.8V so it's not really safe for memory ). Sometimes 12-11-11 but it doesn't change much. On your board you can probably run CR1 on 2 memory sticks. I don't know if you make it work on 4.
 
If you can run 4 memory sticks at about the same clock and timings but CR2 then it will be probably faster than 2 sticks at CR1. Still in XTU you need high memory clock and tight timings to compete with others. Most users run it on samsung b 2x8GB ~3866-4000 12-12-12 ( voltages above 1.8V so it's not really safe for memory ). Sometimes 12-11-11 but it doesn't change much. On your board you can probably run CR1 on 2 memory sticks. I don't know if you make it work on 4.

Ill remove two and see what I can get. Then Ill crank the voltage up and see what we can get on all 4. Ill make a separate profile so that I dont have to worry about that. Ill just load it for running these short term benchmarks.

Lets hope it doesnt go pop. :eek:
 
A word of advice/caution. If you have a spare drive I would suggest you use that. This kind of "play" will ruin your windows install over time. So having it on a "main" drive will just bring you sorrow.
Using a clean OS helps and also setting the memory size limit in Windows will help you push it a bit further
 
A word of advice/caution. If you have a spare drive I would suggest you use that. This kind of "play" will ruin your windows install over time. So having it on a "main" drive will just bring you sorrow.
Using a clean OS helps and also setting the memory size limit in Windows will help you push it a bit further

I actually have three drives in the computer. Ill toss another install of windows on the 500gb SSD as well.
 
Also best to disconnect anything that might get corrupted. If the MBR gets messed up it can take out all the drives.
 
Alright, so have been messing around here. I must have some bunk RAM sticks. Even with voltage set all the way to 1.500 they won't boot at 4000 with only two sticks, and at a 15-15-15 timing. That's with 1.35 on the SA and VCCIO. I also can't boot 5.3 with voltage all the way to 1.55. 5.2 boots no problem and runs XTU at 1.45 volts. But, I'm now getting a lower score than I was previously?? By nearly 90 marks.. my best before was 1677. Now I'm running around 1586..

 
boot at close to the highest speed your computer will boot at and use the asus utility to overclock more in windows. see if that helps you any
 
boot at close to the highest speed your computer will boot at and use the asus utility to overclock more in windows. see if that helps you any
Alright, I'll try that. Currently max it will boot at on 4 sticks is 3600 at 14-14-14-28.

 
Copy that! I'm on it when I get off work in the morning. Also, what are you guys using for a quick screen capture utility? Something that I don't have to worry about using the snipping tool and taking that time.

For win XP or a stripped down OS that doesn't include snip I use GADWIN Printscreen. It's an older program but the one I've always used. Might be something better these days.
 
Yeah that can happen with mem tweaking. There needs to be a balance between speed/timings/voltage plus the CPU portion as in SA IO overvolting is just as bad as under. Look at voltage like a relay race when things are in sync the handoffs are easy when they're not you get errors or a missed handoff which has to wait till the next cycle.
 
Ok. Back at it. And appears the Asus memtweakit doesn't work. It just loads up the start page but won't let me do anything. XTU also won't show RAM for me. Is there another utility that I can try to tweak RAM timing within windows?

 
Iirc None of them are any good and most if not all won't allow you to edit the CAS. Most of us just use BIOS for memory timings. Time consuming I know. Also just an FYI, almost everytime I've borked an OS was while messing with timings. I know you have the extra drives ready so this is just a heads up. :)
 
There is a program I use for screenshots called snapshot. It is really lightweight and will grab a screenshot and save it with a single click. If you follow the benching guide in my Sig I have a download link there as well as directions.
 
I think this RAM is just bunk... Doesn't matter if I'm running 2 sticks or 4. With timings even remotely tight it doesn't like it.

 
Back