Printable View
http://members.lycos.co.uk/csmo/Comp...700/piFast.gif
http://members.lycos.co.uk/csmo/Comp...uperPi_All.gif
I noticed that I cannot clock my memory as high with the 3700+ compared to the CAA2C 3500+, did you guys notice the same thing? My Mushkin memory can do around 260 @ 1.5-2-2-5 with the 3500+ but can only do 246 with the 3700+. :(
I know at least one person which is able to run his TCCD higher than before on Sandy than on a Winny... Colin, probably some of the timing values not quite right... but don't ask me for the correct ones :D Refer to kakaroto's DDR600 guide, I think, he has added some info on TCCD and San Diego memory controller relations there...Quote:
Originally Posted by Colin
pifast with very tight timings 1,5-2-2-3 ,read preamble 4,5
http://upl.silentwhisper.net/uplfold...d1/pifast8.jpg
Cheers, thanks for letting me know. I will definitely check out the guide! :DQuote:
Originally Posted by mrlobber
Finally got my rig up! (FX55 SD)
Hopefully will have some screen shots up once I get my new PS (the one I ordered just didn't have enough... POWA! lol, anyway..
I did manage a nice 3.1ghz very stable (VapoLS), and a 10,000 3dMARK05 before things went wrong >.<
The graphics cards are sucking the most power, I think.
FX 57 has cold boot problem, I am helping setup one on phase change. beyond ~-20c no boot :(
Hopefully I am wrong but right now it looks like a problem for this chip.
It makes no sense that the other 90nm SanDiego's can handle -50c+ but this one can't handle -20c. I wonder if it's just THAT chip? Interested to see results with this thing.
kakaroto had one sd3700+ that could handle -20°c too ;)
CABGE 515SPMW @ 310x10 1.63v Prime stable. It can go higher but I wan't to keep RAM at full capacity. This will be 24x7 setting. Running on MachII GT -27C idle.
hey guys.. jsut wondering whether is there a week number to look out for the san diegos?
Try using a 10.5 multiplier (yes I know it's not actually 10.5) It allowed me to clock my BH6 exactly the same as on my previous 3500s.Quote:
Originally Posted by Colin:I noticed that I cannot clock my memory as high with the 3700+ compared to the CAA2C 3500+, did you guys notice the same thing? My Mushkin memory can do around 260 @ 1.5-2-2-5 with the 3500+ but can only do 246 with the 3700+.
Before I was limited to 240 - now 260.
just what i didn't want to hearQuote:
Originally Posted by FUGGER
anyone know?
is there a good week to look out for? =/
Heh, try to avoid 0516WPAW. Under air it quite sucks. Then had it under dry ice this weekend. Maximum I was able to boot to Windows was 3130 Mhz @ -37C. Built in memtest run also at 3300 Mhz, though (I suspect some problems with my mobo as well, but anyway).Quote:
Originally Posted by babyelf
Btw, any information on which steppings are sent out right now? Haven't seen much 0517 and later San Diegos...
There are better steppings but if you can get them cold there are pretty good .. http://www.xtremesystems.org/forums/...705#post904705 and 3.063Ghz here http://www.xtremesystems.org/forums/...700#post904700
I have a 05-17, about 2750 on 1.45 volts in bios. I did a suicide of @ 2960 at 1.7, I didn't want to do more because of stock HS, hehe. On water I would say 24/7 prob = 2.8, on phase or colder who knows, 3.2-3.5? I'll do some real testing and post it in the offical OC thread with some screenies. Also, the stupid heat spreader is on also, making the chip much hotter than what it should be, I suspect amd is using that plastic goo crap now under the spreader instead of nice thermal paste, I will be performing some sugery when the water gets here, hehe.
Your comment on the goo reminded me of something. I was doing some things while Smartguardian was on the screen, and every task from opening folders or internet to doing Pi, caused the cpu temp to soar momentarily then almost instantly as the task was done returned to normal. I did not know if that was accurate to what was happening or just picking up extra voltage somewhere while the task was being performed.
When under dry ice, I observed the same - temperature jumps at about 10-12 C. At full load, however, the temps remained the same. Weird.Quote:
Originally Posted by coop
i'm getting a san diego.. it's being shipped over to the supplier now.. might get to choose one..
that 0516 WPAW looks nice..
so there's no really a good one to look out for right now?
u think the 0517 will be the same as the venice's 0517 lol.. i doubt it.. hmm
yes try to get a 05-17, every one I have seen has had good results so far.
for san diego? mr. rock lee? :p
so it's the same like the venice?
I have one of the 1st SD 3700's which is week 15 and I am at 2904 mhz @ 1.62 vcore 100%. So I think the earlier version it is the better.
hmm the results are pretty mixed up at the moment.. i'll just have to ask for the list of stepping he have when it arrives.. then choose one that i think will be the best =/
PMW or PAW at least
It would seem so, did u check out the officail oc'ing thread yet? pick a high oc stepping or 2 and hope he has it, heh.Quote:
Originally Posted by babyelf
yea.. but it seems like all the results are abit mixed up =/
i'm hoping for a 3ghz one with my water :p like my venice.. :D
Thats gonna be a tall order since I have only seen one that will do 3ghz on water. I can get to 2970 on water but not 100 stable. 3dmark and what not but it wont prime or SuperPI.Quote:
Originally Posted by babyelf
hehe.. well.. at least let me dream abit right? :P
it's coming soon.. so will be able to choose later on
Week 15 seems to be a great week. I've tried two other weeks (one earlier & one later) that didn't even compare.
thnx for the info BMORIN.. i'll be keeping an eye for it :)
but at least i have to get PMW or PAW.. i think they are the best in the batch
Anyone with this SD stepping.....Never saw any that start with a Letter on serial#......
http://img291.echo.cx/img291/3414/screenshot0741kk.jpg
anyone knows whether this stepping is good or not ? cabge 0516 tpaw.. thx :)
Hi everyone, I would like to start an experiment and need everyone who has a sandeigo's help, with the recent discussion in this thread of sandiego core temps jumping, it would be good if we found out what the heck was going on with these cores. We need to know if A: they are just like that, reguardless of heat spreader, or B: it's proportionally due to the heat spreader. Mainly, this is what I need from you all.
1. Using any form of temp monitoring, ie : smart guardian, temp prob etc. Simply run the 1 minute bench of super pi, make sure u can see the temp go up and down, record your start temp, then record your highest temp during the bench.
2. State whether u have a heat spreader or not.
Your form of cooling does not matter, as the point of this is to see how much the heat spreader makes temps go up while doing somthing fairly intensive, and to see the temp spikes. This will help us understand this core better, and enrichen this thread :) It's not called the officail discussion thread for nothing :P
You merely need to post your results in this thread, name your post , TEMP FOR SANDEIGO, and I will record everyone's temp in this post below.
RESULTS :
SPREADER--------------------------------------------------NON-SPREADER
Quote:
Originally Posted by irenic
Read before in the thread, I've personally seen 2 cpu's with that stepping, both were not able to complete SuperPi 1M on air and stock volts higher than 2500 Mhz. For 2700 Mhz, they needed almost 1.68 in bios.
Bull:banana::banana::banana::banana: ;)Quote:
Originally Posted by mrlobber
http://img66.echo.cx/img66/4887/39449qf.jpg
That is a 0516 TPAW (watercooled)
And i'm running now 24/7 2850mhz 1,45v on boxed air because i'm waiting for some seal string to put my 3700 under my recently bought mach II @r507 :banana:
nice one el snorro.. looks like some good stuff.. :D i'll try my luck with these if i can get them
yay getting a CABGE 0516 TPAW hehehe
wish me luck :D
Hm, I was just sharing my experience. If you have been more lucky, it's your luck, but don't call misfortune of others to be "bull****" as with all my knowledge in English, this word clearly doesn't belong there in this case.Quote:
Originally Posted by El Snorro
I bought 2 of 0516TPAW's, and both were equally bad, one was tested by
my friend also on water, and didn't clock higher.
On dry ice, my particular one maxed out at 3130. Hardly to call it a great overclocker if another of my friends has got a 0515VPMW to 3 Ghz on air :rolleyes:
Now of course, giving my luck of buying cpu's, I guess, a 0515SPMW or like that would happen to be the worst in whole batch for me as well ;) But I think you got the point.
lol mrlobber.. don't put ur hopes too low.. ur next one will be a bomb :D
anyway anyone know whether the serial affects overclock like what steven say aobut the venices?
wow nice speed u got there el snorro :)
btw i guess even with same stepping, it doesnt mean that chip can achieves exactly like the other chip does..
heh u telling who?
if me i know man hehe.. 3 0517 DPMW and only 2 passed 2.9 and only one did 3ghz stable.. the other one about 2.8+ghz
i've seen a few of el snorro's and they were hitting some good speeds.. so it won't kill to hope for the best right? :)
Oh well heres the result....Quote:
Originally Posted by Dumo
http://img156.echo.cx/img156/4863/screenshot0845uo.jpg
I'm inclined to think it is coming from where Smartguardian is reading the voltage from. Doing a 1m Pi the temp goes instantly from 27 to 38 degrees and stays there solid until it is finished, then instantly returns to 27. When I used cold water it would increase 10 to 11 degrees as well, from 11 or 12 to 22 to 23 during a task then return immediatly after completion. That was even during a 16m it was the same.
no its just the cpu puts out close to 200watts of heat and possibly more the heat will go instantly that much higher because the thermal properties of the die, the heatspreader, the TIM, and then your hsf/waterblock....when im running my mobo in a chill box at -15c on a stock hsf, my SD 3700 doesnt just that quickly do to the surrounding air temp. remove the HS if you have a copper hsf/waterblock and it will still jump but what it jumps too wount be as much. you will see a jump on anything from stock hsf to ln2 when a proc puts out that much heat and especially when it still has a HS on...the HSs are junk....Quote:
Originally Posted by coop
I just did a bit of surgery, dropped my idle temps from 42c to 32c literally. Also, I guess I answered my own question, heh, what was on that spreader? Plastic gooey shin etsu, same stuff amd has been coating stock HS's with for yrs, NOT good, 1 layer of plastic + 1 layer of copper + 1 more layer of plastic (for those who are too afraid to use anything else) + a stock HS = crap for temps. I imagine I would have been at about 45c idle if I did have that 2nd layer of junk. So glad I took it off. The core is small, heh. fyi, I had to sand down my dfi HS bracket all the way except for 1/10 of a centimeter. I ran a 8 min super pi just to see how high the temps would spike, they didn't go past 42c. 10c must be pretty consistent for a temp spike seeing as how everyone else is getting about a 10c spike, 3 gigs, here I come, heh.
I am so tempted to remove the HS. Warranty scares me a bit tho. 10c drop or even a 5c drop would help me a lot. I idle at 28c now with wc but my temp is 10c colder then that so Maybe I can get 24c 23c idle. More i think bout it the less likely it would help me much.
Quote:
Originally Posted by Colin
you arent running tccd though. :p:
Quote:
Originally Posted by Crazykooter
yes im very tempted too.not the warranty because AMD are absolute nightmares when it comes to using that(took me several months with a barton to get replacement)but im just too lazy atm and i also like to tighten my RBX without having to worry about crushing the core.
but it will help mate regardless of temps. a mate of mine did it on a good water setup.even though his max load temp was 40C with max overclock of 2.9GHz @ 1.55V when he raised to 1.6V (43C load)he still couldnt get anything above 2.9GHz stable.
after removing the HS he can get 2.93GHz stable at 1.55V and loads at 39C.regardless of temps it seems to raise max overclocks 100-300MHz from the results ive seen or read about.
How hard is it to remove the IHS? I could diffinitely use it since Florida temps suck and my rooms ac sucks too, idle around 40-42 with my 3700+ sd @ 2.7Ghz on 1.55 Vcore
Just bought last night on ebay CABGE 0518 UPCW - that's a 3700+ of course - anyone seen it before ?
takes 5 mins. 5 careful careful mins, heh, sanding down the stock dfi HS bracket took longer, make sure u don't go into the chip deeper than 1/2 centimeter. you'll hit a chip, poof, no chip. mark it on your blade so you know to never go past that point, don't slip, and your good to go. Start with the corners, it's easierQuote:
Originally Posted by uniacid
never seen a UPCW before..... I wonder what part of the world it came from? Let us know how it clocks.Quote:
Originally Posted by S&M
guys.. so there's no early batch chip thingy right? lol.. 0516 TPAW here i come then :p
:D
Definitely not much results with week 0518 and later chips around so give it a try and don't forget to inform us! :DQuote:
Originally Posted by S&M
here is my result with 0516 TPAW.. i can run super PI up to 2.95ghz but still failed prime at 2.8ghz. i tried 1.66vcore the highest because i dont think i need more than that..
hehe same stepping ;) what's your full badge?Quote:
Originally Posted by irenic
ada3700daa5bn
cabge 0516TPAW
1240918d50079
still working on the oc, current 2778@1,5vcore
http://users.telenet.be/duploxxx/oc/...v%2010x278.jpg
i only remember this "cabge 0516 tpaw"
.. quite lazy to reopen my water block.. hehe
btw, urs seems like a good stepping.. mind tell me ur other setting in bios?
So you sand down the IHS 1/2 cm ? wouldn't happen to be any guides I could read up on?Quote:
Originally Posted by fareastgq
I did not sand down the IHS, I sanded down the yellow thing that holds the stock heatsink in place. it bolts onto the MB. U have to sand down the same amt as the width of the IHS so your heatsink will make good contact.Quote:
Originally Posted by uniacid
OH now I understand, well I'm using watercooling and I guess I could lap the cpu block for better contact and tighten it down moreQuote:
Originally Posted by fareastgq
it depends on which hsf/waterblock you are using! with swiftch stuff, you do NOT need to do that...Quote:
Originally Posted by fareastgq
Odd...I ran superpi 1M at 11x260 and got a few hairs of a second fasterQuote:
Originally Posted by irenic
I don't know the week number of my processor, sorry :(. Taking off the block and cleaning off the IHS is a big chore for me :/
Quote:
Originally Posted by Tanktanium
maybe tighter timings on your setup?Drop your Idle Cycle Limit to 0 with A64 tweaker and i bet SuperPi is faster still. ;)
Tanktanium, very similar CPUs?
vcores etc.;)
http://www.geocities.com/pumbertot.1...eenShot010.gif
I'm running on 1.4v vcore :D
Also got a CPU-Z validation of me hitting 11x273 (3005mhz) Not stable, of course :D. 3.8v 2-2-2-5 redline 1:1
BIX3 and PolarFLO TT cpu block. No chilling.
http://valid.x86-secret.com/show_oc.php?id=23557
you mean the mem settings? its all on minimum / tight timings except trc/trfc 8/10 read it from the dfi-street site. for bh-5 it is easy on the dfi, on my asus i have tccd and need to do it with a64 tweaker and that's a pain :slap:Quote:
Originally Posted by irenic
syetem can go faster but need to mod my heatpipe and find some spare time...
duploxxx
for that 290htt, i ran my ram at 200mhz but i forgot to set it back to 2-2-2-5 .. i just ran it at 2.5-4-3-10 since that was the timing i used @ 280mhz ..Quote:
Originally Posted by Tanktanium
Hello,
How about San Diego 3700+/4000+, do they have coldbug issue?
My 3200+ Venice have it, with VapoChill LS it did almost 3256MHz and with DI does'nt even 3.2GHz without freezeing on desktop..
So, how is't with San DIego :confused:
Mine didn't have.Quote:
Originally Posted by Dani
Your SD or Ve?Quote:
Originally Posted by mrlobber
If SD so give me some screenshot :)
How many MHz from Air/H2O -> DI :)
..and stepping and batch number, pleace :rolleyes:
This is the San Diego oc discussion thread :p:Quote:
Originally Posted by Dani
Here: http://www.xtremesystems.org/forums/...146#post911146
CABGE 0516WPAW
Not much more around as that was my first DI run ever. Gained around 500 bootable Mhz in Windows from Thermaltake Sonic Tower + 120mm fan. Saw posting 3350 as well but not able to boot to OS for quite mysterious htt problems.
Regarding Venice, some of my friends put a 3200+ under LN2 this weekend, they had the colbug kicking in at -30 to -60 temps. They still managed some benchable 3.3+ ghz, but had some weird experiences anyway.
Hey guys, I just went through pages of this thread...but it's kinda hard to follow which steppings are the best...
I would like to buy a 3700+ SD, can you give me some steppings which are definitely worth it?
Thanks in advance..
Tim.
Hi guys,
Do you know where Can I have a list with 3700+ San Diego Steppings?
I'm currently building a database :)
Thanks :)
Just got a CABXE 0512XPMW - 3700+ (newegg)
going to use it under phase change (first attempt), I hope it's a good one.
Tim.. try 0516 TPAW..
so far it's looking the best.. my 0516 TPAW is coming if u can wait a while before getting it
Hey i have to pick up 2 3700+ SD's next week, they air both for OC on air. Can you guys advise me a good stepping?
Thanks
I just got a CABXE from the egg. hope its a good one..
Have people been consistently getting higher clocks on BH-5 and TCCD with San Diegos over FX-55's? If so I may have to make the switch :D
DFI ULTRA-D
3700+(0516WPMW)
TWINMOS BH-5(UTT)
TT SONIC TOWER
http://img98.echo.cx/img98/1114/kik7ek.jpg
Is that voltage right?
just picked this up. anyone with same stepping what was your results?
3700 sandy
cabge 0517 upbw
Holy $hit!....I second that...Quote:
Originally Posted by zxlr8
1.4Vcore -> Cpu @ 2880
3.37Vdim -> 1gb BH5 @ 262
Wow.....
In my head Peter Lu is known to get good hw ;)Quote:
Originally Posted by s7e9h3n
Hello everyone, I am new here. I think either my cpu has a weak memory controller, or y motherboard does not support it well yet. I have a CABGE 0517 UPAW 3700+ that I bought retail on an msi neo2 platinum mobo w/ xp90, and g.skill pc4400 tccd ram. I have latest bios by need.for.mhz, which is a 1.8mod that locks in 1t for san diego and other timing tweaks. even before this, I was able to run my old winchester at 280x9, with memory 1:1. Now I have my 3700@250x11=2750mhz @1.6v, with 166 divider to take ram down to 208mhz. 1:1 with ram at 250mhz isnt stable, 275x10 is far from stable :confused: I know the neo2 does not officially support the san diegos yet, and I have seen many people running 300mhz+ 1:1 on san diegos, but on different boards. so 1)is my cpu's memory controller weak? 2)do I just need to wait for a proper bios? or 3) do I need a new motherboard? (would be new dfi nf3) Thanks for any help or suggestions you could offer.
on the quality of 3700s that we can buy in the USA? I am getting the feeling that they are not any better than the POS that I have now (can get stable at 11x250 or 10x270 or 9x300or - you get the idea)
Tried 8 sticks of memory, 2 DFI boards and the best WC you can get and still POS. My luck that I would get teh one bad 0517 :upset: If anyone has a place in the USA that has something resembling something decent, please post here. Otherwise I am going stock and quiet and give up.
I can run 10x274 stable with 1.475 volts. it's all in the sig. exact same stepping, I wouldn't give up hope on it yet until u know for sure.Quote:
Originally Posted by arnemetis
Quote:
Originally Posted by s7e9h3n
that s only 1M. for prime i think he have to set more vcore.
free
Well, I got mine yesterday. It's a CABGE 0516MPMW, and it seems to be utter crap. I'm not able to finish 3dmark 01 at anything over 240x11 (2640), and the most primestable I'm able to get, is 255x10, and this is with 1,65v :(
I only just got my Ultra-D too, so I don't have much experience with it. I'm using bios 6-18 on it now. Any tips? I really really hoped for more :(
Meh, I'm giving this chip up. It's impossible to stabilize it at anything over 2.55GHz. It's so typical that I had to get a crap chip. Dang my luck!
yeah i got luck with my first san diego had so much S754 and winchester cores... all no good but this baby rockZ 3200mhz is reachable and 3300+ suicide.
only one thing i got a messy mem. controller my bh-5 could do 282-2-2-2 but in with the sandiego it reaches max. 269 fully stable. i ll try my a-data tccd soon and let ya know the results.
seEn
try moving you dimms around. use orange, test and then swap the modules into the opposite orange and test. then do the same with the yellow. may or may not help, i gain about 3-5mhz in orange by swapping them to a certain place.Quote:
Originally Posted by andL64
I have a quick question what is it taking most 3700+ SDs at 250x11.0? Voltage wise...
I tried v1.350 + 123% and it wasn't stable...so I am not sure...I really don't want to give it too much being on stock HSF.
1,5V-1,55V or max. 1,6 if it cant reach get better cooling.Quote:
Originally Posted by ugp
seEn
Thanks...I am trying 1.500 + 110% It is actually pretty cool still...idling at 36-37C
Could too much voltage make it unstable?
no, but better temps for higher oc :D
i didnt really gain anything over 1,6-1,65V but this is @ 3200-3300
with 1,52V i get 3100 stable so dont gain much because temps getting worse....
seEn
Well I am sitting on 250x11.0 right now @ v1.55
And here is what I got...
http://xtremesystems.org/forums/atta...id=33354&stc=1
http://xtremesystems.org/forums/atta...id=33355&stc=1
What do you think...and it is a 30 sec 1MB SuperPi
well it worked :D ,thats good, just look here:
http://www.xtremesystems.org/forums/...id=33356&stc=1
this is with 252-2-2-2 very nice i think!
http://www.xtremesystems.org/forums/...id=33359&stc=1
here is spi1M @ 3190mhz with 245-2-2-2 (think so)
http://www.xtremesystems.org/forums/...id=33358&stc=1
here is PC2002 old but very nice just look @ memory and cpu score my mem. score increased very very much only with cpu speed.
http://www.xtremesystems.org/forums/...id=33357&stc=1
3272mhz sandra cpu-bench.....
i need some better cooling its with my modded mach2 (R290) when i get condensers i ll build my first single stage ondie cooling :)
seEn
It wasn't stable....:(\
Going to test 250x9.0 and then maybe 270x9.0...try to max out my RAM..
ok so try 1,5V +113% it should give ya about 1,568V or someting like that.
set ltd bus to 1,4V, and yes set some "ugly" ram divider like 166 for 250x11 so ur ram cant be the problem. i think the chip should at least do 2700-2800 with boxed at 36C
what stepping u got?
and just remember better temps will gain you much.
i also got some pre-tested venice with boxed @ 2,7 but it has the cold bug so i bought the sandiego next day :D have to get rid of this venice so its all bout luck.
seEn
ADA3700DAA5BN
CABGE 0518UPCW
1250228E50772
Um...does the LDT Voltage matter...I was reading that it doesn't...
And the RAM runs fine at 250x9.0 and that is what I had it in with my old DFI...
LDT has only made me unstable in the past, temps not being the problem. just depends on the board and chip i guess. test to find out! :toast:Quote:
Originally Posted by ugp