With HT set to 1.0 how high can the Frequency go without hitting cold bug?
Printable View
With HT set to 1.0 how high can the Frequency go without hitting cold bug?
Ahh, the day after Christmas. When peoples hearts are full of joy and human kindness..
Well, I thought they were till I came here and saw the old Lets bash X and inflame the guys..
Ok, took my shovel and cleaned out the garbage and then did a little sweeping to clean it up.
Gentlemen, be friends will ya?
Man, can't ya just sit back and say lets see what this new beast from AMD will really do in mortal hands?
And then this troll bringing up ancient history about Intel as if the Universe would have changed had they been lilly white and angelic in their business practices.. JEEZ..
If your so riled up that you can't come here without fighting maybe it's time to go release some of that testosterone in another way.
No, I'm not going to say what I'm thinking but for Christ's sake literally, relax a bit will ya?
I did 220htt with -160 °C ht 2.0 1320 MHZ for all benches and found out (somehow) settings short afterwards with dice -60°C core idle 250htt boot - I think HTT will raise sooner or later for the results. Testing with multi is just so easy for initial, short testing and you tend to leave out htt for a second ;) but 300 HTT has already been reached with Destroyer on Air and I don't think -195° will pull it down to 200...
Wrote this in another thread
http://www.xtremesystems.org/forums/...&postcount=285
Sample calculation for total fps, just to show (very simplified):
Resolution = 800x600
C2Q: 10 seconds 30 fps, 60 seconds 50 fps, 30 seconds 100 fps = (10x30 + 60x50 + 30x100) / 100 = 63 FPS
Phenom: 10 seconds 40 fps, 60 seconds 50 fps, 30 seconds 70 fps = (10x30 + 60x50 + 30x100) / 100 = 54 FPS
Resolution = 1920x1200 (video card cuts frames on 50 fps and lowers the rest some)
C2Q: 10 seconds 25 fps, 60 seconds 40 fps, 30 seconds 50 fps = (10x25 + 60x40 + 30x50) / 100 = 41 FPS
Phenom: 10 seconds 35 fps, 60 seconds 40 fps, 30 seconds 50 fps = (10x35 + 60x40 + 30x50) / 100 = 42 FPS
Allways we have some dividers, but way to calculate memory freq. is different on old X2 and Phenoms.
What you are doing in this forum if you don't see different betwen HT and HTT?
I know that you are from usa, but think about you read here, is not that hard.
Phenoms have 2 freqs. one for NB (what i belive that is memory controller, which is said to have to be either equal or higher than the HT freq, on my phenom 9850 it is 10 by default (2000 mhz) and so is HT Link.
any1 who have a phenom can easyly see it in memory section, NB speed, and you got the normal HT link.
Does the NB freq matter at all ?this is though for memory i reccon, but that is what I!! belive, so theese are no facts except what the standard frequencies are. ( what i mean is, does the NB freq make the chip coldbug at all, or is to just to leave it at standard. is this whats been done on all sessions. i already know 1000 mhz HT link means that many chips can go down to almost 190- on LN2 but well, they have never ever mentioned the NB freq.
SOF, tested that at all ?
And yeah true, phenoms have dividers for memory :) pitty i cannot run lower than CL4 on 1066 mhz.
Memspeed = ref HT / mem divider
NB speed = ref HT * NB Multi
CPU speed = ref HT * CPU Multi
HT speed = ref HT * HT multi
HT multi must be less or equal with the NB multi.
The on die memory controller and the L3 cache run at NB speed.
HT speed is most interesting in dual/multi socket systems because the chips are connented via HT.
On single socket systems it may have an minimal impact on games. On boards with sideport memory this memory is connected via HT also if I understood macci correct.
i think sometimes people get confused and say it. i know i do sometimes. maybe they are thinking of hyper transport terminal. idk
and yes nb/ht speeds do matter. you can see there are some pics around here with people messing around with the nb speeds and theres a big difference.
http://valid.canardpc.com/show_oc.php?id=471470 another one came in today.
"ref HT", "HT multi", "HTT", & etc. naming conventions are why I started using "clock" or "reference clock" instead of the other options(though "reference clock" isn't exactly right, either, it's gets the point across) and using "HT Link" helps as well, though it isn't as convenient as a couple of capped letters. HT Link and clock speeds are often confused in all those abbreviations and it's very confusing for n00bs as well as people translating posts to another language. But the labeling discussion has been going on almost as long as the A64's have been out. Too bad things got off on the wrong foot so many years ago - now it's just a mess. :(
Nice! With all the great results pouring in I hope I can find a way to snag one of these come February. I'm already getting unhappy with my 9950 OC results - after looking at all the 940 OC's 3.2GHz just doesn't seem so good anymore ... ;)
Enjoy with this guy benching a phenom II 5,2GHZ, (with video)
http://forum.breakthelimit.net/index...st=0&#entry754