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

My new system OC: ASUS M4A79T Deluxe + Phenom II 945 Deneb

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
velozzity said:
but I'm sure someone will put me in my place and prove me wrong soon.
Click to expand...

Yes it it was Intel ram, but everything else was wrong :)

It is neuro not necro, I do not play with dead flesh. (except cow flesh, and thats not playing thats eating :) )

The Intel kit was Corsair triple channel 1600 MHz 7 CAS kit rated at 1.9v

He managed to pull over 1800 MHz on the RAM while doing 3+ GHz on the northbridge.

So high NB and high speed memory are not mutually exclusive... perhaps maybe just with OCZ? (jk)

AMD can do amazing speeds, but you MUST MUST MUST have stable memory, otherwise everything falls to ****. Much more so than on previous platforms and other architectures :)


1800%20ram.JPG

Personally I have not had a lot of luck with OCZ on any platform I have run them on. They are a great company and their customer service rocks... they just do not make good RAM imho.

Dunno why, they buy the same IC's as others....
 
Velozzity, I am surprised your ram is running that tight! 5-5-5-19 @ 1333mhz! I can't seem to find the thread, but when you run a T1 Command Rate, it cuts every timing down by one. When you run T2 timings, it means there are two ram clocks before the data is registered or read. When you run T1, it means there is only one ram clock so all your timings are dropped one.
 
They suggested 5-5-5-15 T1 for 1333? Doesn't seem right to me... You are talking about ddr2 800mhz timings for almost twice the speed.
 
So I should lower my RAM clock and just tighten up it sounds like...
Man I dunno how these guys figure out the timings anymore... I used to spend DAYS (full weekends) to tweak BH5 ram. It's just not worth it for me... or perhaps there's an easier way.
 
I understand that, but I don't think you understand me. Right now, this exact moment in time, you are running 5-5-5-19 @ 1333mhz. When you run 6-6-6-20 @ T1 is is actually 5-5-5-19 where as 6-6-6-20 @ T2 is 6-6-6-20. SO you are indeed running the suggested settings.
 
Not trying to run away with a thread started to help someone with his overclock but this is a quote out of the link I posted above

"Update May16th 1:25AM...5-5-5-24 1t hits back. you guys looking for 1800MHZ ram at 8-8-8- or 9-9-9- forget it, with the NB around 2400 to 2800MHZ 5-5-5-24 1t 1333 is where you need to be."

Notice he (tony of ocz) said 1t, so by your rule the timings are actually 4-4-4-23.
 
ChanceCoats123 said:
I understand that, but I don't think you understand me. Right now, this exact moment in time, you are running 5-5-5-19 @ 1333mhz. When you run 6-6-6-20 @ T1 is is actually 5-5-5-19 where as 6-6-6-20 @ T2 is 6-6-6-20. SO you are indeed running the suggested settings.
Click to expand...

Actually it is more the other way around, 6-6-6@2t is more like 7-7-7.

The CPUNB speed seriously hampers the effectiveness of running 1600, even 1333 is limited with a stock CPUNB. In my own testing(mostly games) there was no real difference between 1600 and 1333.
 
ChanceCoats123 said:
I understand that, but I don't think you understand me. Right now, this exact moment in time, you are running 5-5-5-19 @ 1333mhz. When you run 6-6-6-20 @ T1 is is actually 5-5-5-19 where as 6-6-6-20 @ T2 is 6-6-6-20. SO you are indeed running the suggested settings.
Click to expand...

velozzity said:
Not trying to run away with a thread started to help someone with his overclock but this is a quote out of the link I posted above

"Update May16th 1:25AM...5-5-5-24 1t hits back. you guys looking for 1800MHZ ram at 8-8-8- or 9-9-9- forget it, with the NB around 2400 to 2800MHZ 5-5-5-24 1t 1333 is where you need to be."

Notice he (tony of ocz) said 1t, so by your rule the timings are actually 4-4-4-23.
Click to expand...

Guys, carry on with the conversation b/c this is also helpful with me, but what are you guys talking about that one timing is actually another... are you comparing 1T timings to 2T, or are you saying something else... you guys mentioned...
"6-6-6-20 @ T1 is is actually 5-5-5-19"
"5-5-5-24 1t............ so by your rule the timings are actually 4-4-4-23"

*Edit*
Okay, so I read the link...
I guess that 1333 6-6-6 is similar to 1600 7-7-7, 1333 5-5-5 similar to 1600 6-6-6 and so on? One clock tick tighter and dropping one divider is what I should be doing...
I wonder if my OCZ Gold (rated 8-8-8) can run 6-6-6 @ 1333... I sure hope so!
Still, the quote in WHITE BOLD above makes no sense to me.
 
Last edited:
Chance, I am not disagreeing with your way of explaining the command rate of 1t and 2t and It's effect on all other timings, only trying to say that at the ocz forums the recommendation was made to set timings at 5-5-5 with 1t command rate to get better than 1800 mhz performance at 1333, I would assume that to mean setting cl-trcd-trp to 5-5-5 and the command rate to 1, If it ends being same as 4-4-4- 2t so be it I am not an computer engineer just another average joe that understood his recommendation to be exactly as stated and no that I had to set it as 6-6-6 and that the 1t rate would make it 5-5-5 effectively.

Also per my comment on 1600+ being difficult and someone to prove me wrong I knew that a few well known overclockers have made it to slightly over 1800 but I still believe that out of 100 sets of memory that are rated at as 1600 mhz, the majority will not overclock and a large percentage wont even do 1600 without a large amount of tweaking, it's not as plug and play as ddr2 1066 for instance. I am stating what I feel is the norm on the street per se not what the statistical exception is.
 
Last edited:
I know what you are saying. I am just a bit surprised that the ram is running that tight to begin with! And then on top of that he wants you to run 5-5-5-24 @ T1 timings. That is pretty damn tight. I can't get my ram to run T1 even with the voltage maxed and the settings all at stock (5-5-5-18-22-6-3). But I also can't run cas 4 @ T2 timings either which is why it leads me to believe T1 is a clock shorter.
 
Godm@n,

What I meant to say and should have clarified is that by Chance's rule as I called it of how setting command rate to 1t affects timings that 5-5-5-24 1t is actually 4-4-4-23 2t, should have typed the 2t so It would have made sense.

Also I am not at the end of the road with this memory, I have actually dropped timings to 6-6-5-17-27 1t and it has so far ran 10 passes of memtest (about 5.5 hours) with 0 errors. I cannot get the 5-5-5 nor 6-5-5 timings to work but bandwidth does go up from 9700 to 9800 and latency does drop from 44.7 to 44.4 just from this timing drop. This also Is with 1.61v, memory is rated at 1.65 but strangely wont work properly any higher.
 
velozzity said:
Godm@n,

What I meant to say and should have clarified is that by Chance's rule as I called it of how setting command rate to 1t affects timings that 5-5-5-24 1t is actually 4-4-4-23 2t, should have typed the 2t so It would have made sense.
Click to expand...

Thanks for the clarification, but where did you get this info? You're saying they are IDENTICAL or are you saying that the performance of both of those settings is virtually identical?? TIA.
 
godm@n, I am neither agreeing nor disagreeing with chance I am only saying that by the way he explains it 5-5-5-24 1t is equal to 4-4-4-23 same as how in this post. I added in the 2t part because I assumed he was talking about 2t because If command rate is not set to 1t it has to be set to 2t.

I understand that, but I don't think you understand me. Right now, this exact moment in time, you are running 5-5-5-19 @ 1333mhz. When you run 6-6-6-20 @ T1 is is actually 5-5-5-19 where as 6-6-6-20 @ T2 is 6-6-6-20. SO you are indeed running the suggested settings.
Click to expand...

he said 6-6-6-20 1t is the same as 5-5-5-19 or at least that is how I understood it, I will wait for him to explain if he meant 5-5-5-19 2t or not?

Be really who cares about the specifics I was just trying to repeat what he said but in reference to timings that the article said use (5-5-5) run the tightest timings that you can and we all know that 1t although harder to run increases performance...end of story.

by the way did you ever get 6-6-6 to run?????

Update, I found this http://www.tomshardware.com/forum/176007-30-dram-command-rate-explained by googling but dont know how accurate it is but, it seems to me it's like an analogy of how long it takes a stock picker on a forklift to go from loading dock to warehouse section (warehouse section being the memory bank and speed to get to warehouse being command rate), the other timings being how long to drive down a certain row, raise forks to certain level in steel, and pull good from location (as in where data is stored and how fast you can get to its exact location with each timing referring to row and column etc.), so by this analogy and post command rate does not directly affect other timings it is just a different wait time per se to do a distinct operation. Now I have really confused myself and everyone with this analogy hahahah
 
Last edited:
For some really odd reason I cannot remember for the life of me where I heard/read that, but it was a few days ago so I think my memory is correct.

I believe it is saying the actual timings are equal to a whole clock lower when you run T1 over T2 instead of the performance being equal.
 
Command Rate... Rate deals specifically in time. The Command rate does not alter the speed of the ram by changing the timings.

"The Command Rate is the time needed between the chip select signal and when commands can be issued to the RAM module IC"
~straight from AMD

Therefore by limiting the time to 1 clock (1T) you are forcing the commands to be issued faster than at a 2 clock (2T) rate. This will provide a performance increase, due to the decrease in time allowed... it is not due to a physical change of the memory timings by the command rate setting.

So 1T will give you the performance of tighter timings, not 'actual' tighter timings.
 
You must log in or register to reply here.

Similar threads

S
My old 7 years system won't power up. Need help
Replies
6
Views
803
Salamandrix
S
heavy666
Using asgard 32gb Loki mems
Replies
2
Views
752
heavy666
heavy666
mackerel
1Rx16 to 2Rx8, filling in 1Rx8 testing?
Replies
2
Views
2K
mackerel
mackerel
N
Noob needs help with old rig
Replies
2
Views
758
EarthDog
EarthDog
funsoul
Need build oc help (5900x, 3600 b-die, 3090ti)
Replies
16
Views
2K
funsoul
funsoul
Back