*NEW* 1800+DLT3C 'B' results! (week 033X, any of the *many* steppings)

[felinusz]

Due to all the new T-Bred 'B' 1800+ steppings recently appearing (KIUHB, NIUHB, JIXHB, JIXIB, and the *still-being-produced*, and legendary, JIUHBs to name a few) I would really like to see what kind of results everyone who has one is getting with theirs! :)

We are looking at week (year = 2003) 03 (week of 2003 = 30 + X ) 3X (hence 033X) chips - these are the newest ones around - as well as the reappearing "old stepping" JIUHBs that I have seen *EVERYWHERE* (all my local computer stores now have them!) in week 0334 batches. Note that so far all of these "new" steppings have appeared only in retail boxes as of yet (the shiny new ones with all the plastic). Most people are getting them from newegg.com - most of which seem to be NIUHB 0334s so far. I got mine from a local store.

I'll start with my just-opened JIXIB 0332:

Does (200 X 12) 2400 MHz @ 1.725 Vcore, 1/2 hour prime95 stable (I'm in a rish to get higher, and I'm not stopping to check for true 24 hour stability at lower overclocks just yet).

Is currently @ (200 X 12.5) 2500 MHz, 1.85 Vcore, 1 hour prime95 stable (so far!)

hasn't been burnt-in yet - RAM is limiting my going higher as well (DDR 266). And my Vcore mod is preventing me from seeing how far it will go on stock voltage... *blah*

I would like to see how each of the "new" steppings is doing with different equiptment and cooling - if you post your awesomazing new overclock here please include your Vcore, and your Multi X FSB ;)

Hopefully I will soon be able to compose a graph of sorts comparing all the new steppings in different weeks, with different cooling. :)

Hopefully we will also be able to compose a list of "new" stepping codes and the associated weeks on them as well ;)

[Lithan]

NIUHB 0334 DLT3C 1800+ Retail Newegg
1.95vcore (1.84-1.90v actual)
12x217
Watercooling
530watt fort
NF7-s 2.0

100% prime stable.

[redgtxdi]

I've read & read & obviously there are some opinions but perhaps somebody knows for sure..........

I have 2 1800's....

#1.) JIXHB 0329

#2.) JIUHB 0323

Is one or the other a better stepping? TIA

[RedDawg41]

You might have to actually test them out on your system to say for sure which one is better... JIXHB seem to be the newer stepping, but I have an JIUHB 0319 that does 2.535GHz on 1.95vcore. you might find one that does 2.500GHz on 1.85vcore which is, imho, way better voltage wise, and not much of a performance difference by 35MHz.

[julianba]

My 1800+ niuhb 0334 dlt3c from newegg is currently running 180*14=2520 mhz @ 1.82-1.84v , it's been running prime 95 for about 45mins. :burn:

It went about 2150 at 1.50v

[wannaoc]

I had a JIUHB 0311 and it was horrible on my NF7-S. I couldn't get it over 2.2 @ any volt and barely to 200 FSB.

[USAPGAPRO]

wow, these new 1800+ are killing the 1700+DLT3C's. I might have to pick one of these up.

[felinusz]

I'm currently burning in at 2.0 Vcore, 650 MHz and 28 DC load.

I ran 2500 MHz stable for two hours at 1.825 Vcore - but it made an error shortly thereafter. 1.85 Vcore makes a overnight (10 hour) prime95 stable 2500 MHz. Hopefully burn-in will help it a bit.

I am very curious to see how that JIXHB does - I haven't seen many results from those - let alone on water.

Does anyone else have a JIXIB, or a KIUHB?

[Arkaine23]

1800+ DLT3C JIUHB 0334

190x12.5 = 2375 mhz at 1.75v and taking my time going up, but there's still headroom

[rommelrommel]

I've got several entries:

0308 JIUHB 2.4 @ 1.8 Vcore (kinda forgot to test this one out, think it has a bit more)

0310 JIUHB 2.4 @ 1.8 Vcore (more vcore makes it hang)

0332 JIXIB 2.45 @ 1.8 Vcore (more vcore makes it hang)

This is with:

NF7-S Rev 2.0
Antec TruePower 550
SLK-800 92mm Panaflo med

I'm trying to find a NIUHB at the moment, in anyone in Vancouver B.C. has seen any let me know :)

[wannaoc]

Looks like my 0311 is the only bad week code so far. Thats good to know.

[brennan77]

Originally posted by wannaoc
Looks like my 0311 is the only bad week code so far. Thats good to know.

Don't use that as a generalization. I have a 311 and there are some others around here that all do 2400 or better with 1.85v. So I think you got a bunk chip.

[v8vega350]

1800+ nuicb 0334 retail from new egg

12.5x208 1.93v in bios any more and i get reboots
36-44c depending how hot it is outside rest of specs are in sig

[felinusz]

v8vega35, rommelrommel - isn't there some kind of pin mod/pencil mod trick you can do with the NF7-S to make it stable with higher Vcores?

[julianba]

Testing at 2534 on 1.82-1.84v, I'm happy with 2500mhz + but why not push it alittle more?:D

[xhoangx]

just got my retail jiuhb 0329 currently at 200x11.5=2305.2mhz at 1.8 vcore, i had it at 12x210 and it posted but then i upped it the fsb to 215 and it wouldn't post, now it won't even post at 12x210 at 1.8 anymore. haven't tried with anything under 200fsb or anythiing above 1.8 vcore. i'm on air by the way.

[iceman2g]

Originally posted by felinusz
v8vega35, rommelrommel - isn't there some kind of pin mod/pencil mod trick you can do with the NF7-S to make it stable with higher Vcores?

I've never heard about such a mod but if this is so the case i'd like to hear a lil about it.

[v8vega350]

i wish there was
this would be a magic chip if there was my 12v line starts getting low around 2300 mhz even on the 400 watt psu i have seen 11.37v at its worst my 12v wires are getting warm comming of my psu also.

it was fine with my 2100+ at 2525 and 2.14vcore it still had 12v on that line but these dlt3c chips are killing it

[felinusz]

I will search around for you guys a little bit - I am *SURE* that someone mentioned it. I believe that you do a pin mod that changes the default FSB from 133 to 166 - and for some reason this will allow the chip to run stable at higher Vcores. Won't help a dippy 12V line though...

I'll be back ;)

[Soja]

These chips are looking good. So 1.85v is the max vcore you can use? That chip is begging for some juice :D

[felinusz]

I can scale mine up to a 2.1 Vcore stably... Then again - I have a different motherboard, and I am not taking advantage of the higher Vcore, I'm haven't tried clocking the chip higher than 2500 MHz due to RAM.

As for the NF7-S mod I mentioned - it is stickied in the Abit AMD motherboards forum (L12 mod) - I am almost *certain* that it is supposed to allow for a higher attainable FSB as well as allowing for more high Vcore stability.

[xhoangx]

i primed my jiuhb 0329 12 hours overnight at 200x11.5=2305mhz at 1.8vcore, now i'm running it at 12.5x200=2505mhz at 1.825vcore, i won't up the fsb any higher until i get some better ram, maybe buffalo pc3700, and i can't up the multiplier anymore because 12.5 is the max on my mb. i have generic pc3200 ram right now and i get a couple of errors at the end of test 5 of memtest, but only test 5. i'm gonna prime this overnight and start trying to lower the vcore tomorrow.

[iceman2g]

Originally posted by felinusz
I will search around for you guys a little bit - I am *SURE* that someone mentioned it. I believe that you do a pin mod that changes the default FSB from 133 to 166 - and for some reason this will allow the chip to run stable at higher Vcores. Won't help a dippy 12V line though...

I'll be back ;)

I think the mod your talking about is the one that got 266mhz cpus to run higher fsb on Nforce2 ultra mobos. I tried it but it didn't help me much it, granted my fsb was already at 233mhz. Vdimm mod would would have gotten me to 240mhz in DC though.

[iceman2g]

Oh almost for got link. Is this mod (http://www.xtremesystems.org/forums/showthread.php?s=&threadid=15300) you were reffering to felinusz?

[SakabaX]

1800+ DUT3C

JIXIB 0339 SPHW

from SVC

Rock solid 24 hours of Prime 2208mhz (192 x 11.5) @ 1.675 volts (1.62-1.65 according to MBM5)
Sadly, Prime craps out after an hour at 2300mhz (200 x 11.5) @ 1.85volts (don't want to put any more juice into the poor thing) Maybe my 350watt Enermax has finally reached its limits.

Strangly, if I try to change multipliers, the rig will not post. (Using NF7-S, had no problem changing multipliers on old 1800+)
Perhaps new chips are multiplier locked somehow? :D

[LBJGH]

I was thinking about getting a jixib 1800+ but after seeing some of the results I've decided to stick with my trusty air-cooled 1700+

[Chupacabras]

1800+ NIUHB0334 (DUT3C)

I decided to see what all the fuss was about--this replaced my 2100+AIUHB0304 Seems to have hit the wall at ~2450 mhz @ 1.87v :(. With all the guys with 2500+ on air, this was a little disppointing, but nonetheless much better than my 2100 which could only get 2200mhz@1.96 stably.

The system boots at 12.5*200 @1.85 but cant run sandra. I tried it at 1.96V, and it ran , but I didnt get a chance to test prime before the heatsink feel off and the system made a huge crackle that scared the sh*t out of me. I'm leary to try the voltage that high. Is the running the 1800+@1.95 less safe than running the 2100+ at that voltage?

Cooling is a Volcano 7+ in a Antec Sonata case. I passed up a lot of week 32 ish JIUHB's for this 34 NIUHB--stupid thing to do, or its just luck of the draw anyway? My motherboard (MSI KT3 ULTRA2-R) is hard volt modded with a resistor, so I can get higher voltages. You guys think this could somehow be having a negative bearing on stability?

BIOS sees the chip at a 2600+ when it's running 196*12.5, that's probably more annoying than the 50 more mhz for me, it would be sweet if it recognized it as a 3000+ or better :)

[felinusz]

I would get better cooling, and try and mount it in a manner that will not be having it falling off! :eek:]

I would be careful at voltages above 1.9 with these chips - but if you are keeping temperatures below 50 then you may as well step upwards a little bit. Go up in little .025 steps to see how negatively it effects your temperatures first.

[G|-|oST]

OK, and not to start any endless arguments, but the whole huge deal about a T-Bred doing 2500MHz is not THAT exciting anymore is it? If, according to hitechjb1's comparison, a Barton operating at ~ 2370MHz is about the same as T-Bred at 2500. So one could easily get any of the recent Barton 2500+'s and run it at that speed and get roughly the same performance right?

[Chupacabras]

That's a good idea.. not falling off is good.

Would a slk-700 be an upgrade? I'm not so keen on spending what it takes to get a slk800, they are very expensive up in Canada, so unless someone else knows where to get them at decent prices, I'm going to stick with cheaper options.

And G|-|oST, while that may be true, you can almost get two tbreds for the price of that barton. Might as well share the wealth ;)

[v8vega350]

Originally posted by Chupacabras
1800+ NIUHB0334 (DUT3C)

I decided to see what all the fuss was about--this replaced my 2100+AIUHB0304 Seems to have hit the wall at ~2450 mhz @ 1.87v :(. With all the guys with 2500+ on air, this was a little disppointing, but nonetheless much better than my 2100 which could only get 2200mhz@1.96 stably.

The system boots at 12.5*200 @1.85 but cant run sandra. I tried it at 1.96V, and it ran , but I didnt get a chance to test prime before the heatsink feel off and the system made a huge crackle that scared the sh*t out of me. I'm leary to try the voltage that high. Is the running the 1800+@1.95 less safe than running the 2100+ at that voltage?

Cooling is a Volcano 7+ in a Antec Sonata case. I passed up a lot of week 32 ish JIUHB's for this 34 NIUHB--stupid thing to do, or its just luck of the draw anyway? My motherboard (MSI KT3 ULTRA2-R) is hard volt modded with a resistor, so I can get higher voltages. You guys think this could somehow be having a negative bearing on stability?

BIOS sees the chip at a 2600+ when it's running 196*12.5, that's probably more annoying than the 50 more mhz for me, it would be sweet if it recognized it as a 3000+ or better :) i think why you never got very far with this chip is because it is a dut3c not a dlt3c

[Chupacabras]

whoops that was a typo on my part..

[Kekkai]

What kind of ram would you guys recommend for someone who wants to get this CPU, with an 8rda or a7n8xix?

I'm looking for best bang for the buck (as cheap as possible without foregoing too much quality).

[wannaoc]

OK, and not to start any endless arguments, but the whole huge deal about a T-Bred doing 2500MHz is not THAT exciting anymore is it? If, according to hitechjb1's comparison, a Barton operating at ~ 2370MHz is about the same as T-Bred at 2500. So one could easily get any of the recent Barton 2500+'s and run it at that speed and get roughly the same performance right?

I would think it still is considering the stock speed they are coming from and that a 2.5G AMD chip is one damn fast chip. Also consider that alot of people still don't get these speeds.

Kekkai:

I would go with any of the listed chips in this thread. There is one for sale retail in the classifieds right now for 65.00 which isn't too bad considering you know the stepping already.

[G|-|oST]

Originally posted by Chupacabras

And G|-|oST, while that may be true, you can almost get two tbreds for the price of that barton. Might as well share the wealth ;)

Two T-Breds for the price of a Barton??? May be a year ago that was the case, but not anymore. What I see in terms of price, actually is very different.

I am in Canada, and right now bigfootcomputers.com sells JIXIB 1800s for $109 Canadian, and the Barton 2500s are around $128 here in town.

$19 difference.

[nealric]

at newegg, 1800s are 51, 2500s are 91.

[G|-|oST]

Yes, but newegg does not guarantee the 'special' stepping of the 1800+, so you don't know what you would be getting. I meant the JIXIB.

[felinusz]

I have said this many times, and will again - I got my week 033X 1800+ processor (in sig) by looking around at local stores and asking to see the steppings on their 1800+ chips before buying. I got my procesor for $82 (canadian). If you have some spare time do this, and save yourself money, instead of buying from bigfoot (great store might I add!). These chips are actually *COMMON* now - my tech teacher bought a cheapo rig for himself, and had me put it together. And guess what greeted me? A JIXIB 0332 1800+ that he picked up from a local by-the-roadside computer store for $78 (canadian). If you look, you will find.

Chupacabras - go for some Winbond BH-5 (or buy the best - OCZ!! - newly converted fanboy right here!! ;)). Check out the memory forum for a sticky with this weeks prices - and where to buy. And get an ABIT NF7-S, not an EPOX or ASUS nForce 2 board ;) :)

G|-|ost - From what I have seen all the newer (week 033X) DLT3C 1800+ processors have been doing fairly well. The NIUHBs, newer (033X) JIUHBs, and JIXIBs all seem to perform roughly the same with decent hardware, and enough Vcore. I wouldn't pass up on a 033X JIUHB over a more pricy JIXIB or NIUHB - if you can find one locally for cheap then buy it!

Some people will always, sadly, have bad luck and get a cludge chip (luck of the draw I guess... :() - so always take any single person's "BAD DLT3C NIUHB" thread with a grain of salt :).

[dryars]

My DLT3C JIXHB 0329 can do 2.43GHz @ 1.85v(bios). I've booted up to 2.5 but windows was damn unstable. Right now I run at 2.35Ghz because of my craptacular temps at 2.4+. An extra 50 MHz isnt worth the +6c temp increase.

[KfistoRok]

Looks like its time for a cpu upgrade... :D

[Sir Barton]

my 1800+ JIUHB0321XPMW does 220x10.5 on 1.66V. anything more than that...it hits a brick wall no matter what voltage i throw at it.

[felinusz]

Have you tried running with a lower FSB? It may be a symptom of your chipset maxing out with it's walled FSB. I see that you are using an Epox board - can you raise the chipset voltage with it?

Is it prime95 stable at 1.66 Vcore (220 X 10.5)? Try for (200 X 12) 2400 Mhz without such a high FSB, and see what you can do. Then try and find a good MHz/FSB balance for the best performance! :)

[Sir Barton]

Originally posted by felinusz
Have you tried running with a lower FSB? It may be a symptom of your chipset maxing out with it's walled FSB. I see that you are using an Epox board - can you raise the chipset voltage with it?

Is it prime95 stable at 1.66 Vcore (220 X 10.5)? Try for (200 X 12) 2400 Mhz without such a high FSB, and see what you can do. Then try and find a good MHz/FSB balance for the best performance! :)

yep ive backed down my FSB to 200 and tried multis 12 and 12.5 at up to 2.0V and still a no go :( but heck, im happy with an 800MHz OC. 10 hours prime stable at 220x10.5 1.66V :) it will POST at 200x12.5, but windows wont boot.

EDIT: yep, i can push my VDD up to 2.0V with this board.

[Kekkai]

guys i just ordered a 1800+ from newegg. I'm upgrading from a 1600 and obviously the only reason why i went out of my way to buy a new cpu, new mobo, new ram to go from 1600 to 1800 was to overclock the s****za out of the 1800+

but i just got a week39 DUT3C Jixib

can't express how dissapointed I am. Should I even bother overclocking this? I'm thinking of ebaying it and ordering another one. None of the local shops in San Diego have any 1800+'s so I dunno what to do...

I'm still waiting for my SLK900a so i cant test it out yet...

what do you guys think?:(

[Hughhefner]

Well from my sig you can see the results of my 5 NIUHB 0334 1800's. They all hit 2365 to 2450 with 1.875-1.9V

Very happy indeed. I bought a bunch of them RETAIL and was able to let 3 go in the classifieds to other LOYAL overclockers.

Man I am impressed with my SWIFTECH 462-V, I put a 120mm to 92mm adapter on it hooked up a 130cfm fan thats actually QUIET compared to me TTSM2, temps are 123F at load 1.9V 2450mhz---WOW!!!

[Kekkai]

I can get my hands on the DLT3C JIXIB week 29, should I get it, or shop around some more for a Week 33+?

[felinusz]

Kekkai - go for it!

I am sure a week 29 would perform just as well as the more commonly seen JIXIB 0332s. Remember that it is a huge generalization saying that the newer chip is alway better - some people have incredible results from some older chips (like the infamous JIUHB 0319), but in general a newer chip is better than an older chip of the same kind. A 3 week age difference is probably negligible though, so I would be happy for the find and grab it. If it is local, and if the price is right then buy it fast! :) :)

As for that DUT3C - bad luck that. I would ebay it or sell it ont he forums to someone who isn't keen on clocking it really high, and doesn't care that it isn't a DLT3C.

Hughhefner - good idea that, helps others out too. If I can get enough flow to do it I will buy a few JIXIBs retail at my local store and let them go @ near cost.

A very interesting aside is that we are currently in something like the 41st week of 2003 - and I haven't seen anyone with a week 40 1800+ TBred processor. In fact the newest processor I have seen someone with is a week 36. I really hope that AMD hasn't discountinued them already (I heard a while bakc that AMD was planning to)! That would really be a sad day for overclockers everywhere... :(

[Hughhefner]

I read in another thread that ALL Tbreds have been discontinued--all they make now are Bartons and thorntons aka Bartons with HALF the cache removed by cutting one of the bridges. Makes sense to only produce 1 tbred style since the 64's are now out.

[felinusz]

Wow. So these chips are the last TBred's we shall ever see. And AMD has taken to locking the multiplier on all the new Barton's too in a very wise move to make more profits.

The AMD overclocking scene is going to become very different very soon.

[KIAman]

What happened to the good ole xp1700? I still got my xp1700+ JIUHB XPMW 0310 running at 12.5x200@1.775v. I think I got this chip a long time ago too.

[dryars]

Is that Air cooled or Water cooled?

[KIAman]

Was running air cooled for couple of months until i recently went to water cooling. Maybe I should push this baby a little more since I got this water cooling.

[AZN]

just got a a dlt3c 1800 week 34 from Hughhefner and the best i could do the first day was 2.4ghz 12x200 @ 1.8v bios on a nf7-s v2 board cooled by slk900u with 92mm fan at about 3200rpm. no matter how high the voltage is set i can get 2.5ghz stable. i've tried up to 1.93v in bios but temps gets out of hand. right now at 2.4ghz while primming its reading 57c so its getting pretty toasty. this doesnt seem at all any better then the 1700s or maybe its just my setup. :shrug:

[hitechjb1]

Tested two Tbred B DLT3C, both do 2.57 GHz with a TT SFII on a SLK-947U
- 1700+ JIUHB WPMW 0310, 2.57 GHz at 1.95 V
- 1800+ JIXIB MPMW 0332, 2.57 GHz at 1.90 V
Temperature at or below 50 C (prime95) in both cases with different fans.

I found that 80 mm fan on SLK-900U/947U performs better than 92 mm fan
- TT SFII 80 mm fan at 4800 rpm (~76 CFM) overclocked 200+ MHz higher than 92 mm Enermax fan at 2400 rpm (~64 CFM)
- Tornado 80 mm fan at full speed (~84 CFM) cooled CPU 4-5 C better than Tornado 92 mm fan at full speed (~119 CFM)

Explanations and details are in the following links:

Testing 4 fans w/ a SLK-947U on a Tbred B 1700+ DLT3C (http://forum.oc-forums.com/vb/showthread.php?s=&postid=2240172#post2240172) (page 16)

Comparing Tornado 80mm, Tornado 92mm and TT SFII (second test data) (http://forum.oc-forums.com/vb/showthread.php?s=&postid=2249967#post2249967) (page 17)

[G|-|oST]

Hitechjb1

Did you get a chance to see what sort of a difference you got by switching from the SLK-800 to the gy-normous 947?

I was considering the switch but figured there wouldn't be enough difference.

Could you please shine some light on the subject?

[hitechjb1]

Originally posted by G|-|oST
Hitechjb1

Did you get a chance to see what sort of a difference you got by switching from the SLK-800 to the gy-normous 947?

I was considering the switch but figured there wouldn't be enough difference.

Could you please shine some light on the subject?


I have not tested head to head a SLK-800U w/ a SLK-947U, so no experiment data.

But I think both should perform almost identical. In theory, if based on the thermal resistance,
- 800 is considered to be 0.23 C/W
- 900/947 is considered to be 0.22 C/W
The fin and copper structure of all of them are very similar, so are their flow resistance characteristic subjected to the air flow of the same fan.

So between the 800 and 900/947 at 2.4-2.5 GHz level of overclocking, consuming 120-135 W power estimated, the temperature difference would be less than 1.3 C, using the same fan. The 1.3 C in temperature would translate into 0.5% difference in CPU frequency which is about 12 MHz.

So the difference is very small and can easily be overriden by fan CFM, the kind of fan used, system ambinent temperature and case cooling. That is not worth changing from SLK-800 to SLK-900/947.


I got the SLK-947U because I needed a new HS, and it was on sale ($32) and cheaper than the SLK-800U/900U. Also I thought it can take a 92 mm fan which would either deliver better temperature or quieter at same CFM. BUT apparently, the 92 mm fan is not the choice for the SLK-9xx heat sink for the reason I explained and showed in my tests - 92 mm fans performs worse than 80 mm fans.

Testing 4 fans w/ a SLK-947U on a Tbred B 1700+ DLT3C (http://forum.oc-forums.com/vb/showthread.php?s=&postid=2240172#post2240172) (page 16)

Comparing Tornado 80mm, Tornado 92mm and TT SFII (second test data) (http://forum.oc-forums.com/vb/showthread.php?s=&postid=2249967#post2249967) (page 17)

[G|-|oST]

OK thanks very much. This confirms my guesstimate on the subject.

[wannaoc]

just got a a dlt3c 1800 week 34 from Hughhefner and the best i could do the first day was 2.4ghz 12x200 @ 1.8v bios on a nf7-s v2 board cooled by slk900u with 92mm fan at about 3200rpm. no matter how high the voltage is set i can get 2.5ghz stable. i've tried up to 1.93v in bios but temps gets out of hand. right now at 2.4ghz while primming its reading 57c so its getting pretty toasty. this doesnt seem at all any better then the 1700s or maybe its just my setup. :shrug:

Oh AZN thats not what I wanted to hear! I haven't got mine setup yet and its the same thing so hopefully it will do a little better.

[Hughhefner]

The temps that AZN has are a bit high to be able to push the processor past 2400mhz.

My rigs that can hit 2.5 "not prime stable-just internet-SETI-outlook" are using a 120mm 125cfm fan necked down to 92mm on a swiftech 462-v, my prime temps at 2426 with 1.9V are "50C",
big difference in temps versus AZN 57C with only 1.8V

[felinusz]

hitechjb1 - One thing I noticed when comparing my older 1700+ JIUHB DLT3C 0319 with my newer 1800+ JIXIB DLT3C 0332 was that the temperature with the same cooling was the roughly the same - although my JIXIB can do *any* overclock my JIUHB could with a Vcore 0.05 lower.

Why would this be?

JIXIB @ 12.5 X 200 = 2500 MHz, 1.85 Vcore, 37 Degrees Celcius Load
JIUHB @ 12.5 X 200 = 2500 Mhz, 1.9 Vcore, 38 Degrees Celcius Load

Admittadly the Arctic Silver Three on the JIUHB had been there for about 3 weeks of 24/7 use, and had probably set - and the JIXIB's arctic Silver was freshly applied - but that shouldn't account for such a small temperature difference at a signifigantly lower Vcore.

Also note that my board at the time, an ASUS A7N8X-Deluxe 2.0 couldn't hop a stable FSB above 203 with a multi higher than 11, and 208 with a multi of 11.

[hitechjb1]

Originally posted by felinusz
hitechjb1 - One thing I noticed when comparing my older 1700+ JIUHB DLT3C 0319 with my newer 1800+ JIXIB DLT3C 0332 was that the temperature with the same cooling was the roughly the same - although my JIXIB can do *any* overclock my JIUHB could with a Vcore 0.05 lower.

Why would this be?

JIXIB @ 12.5 X 200 = 2500 MHz, 1.85 Vcore, 37 Degrees Celcius Load
JIUHB @ 12.5 X 200 = 2500 Mhz, 1.9 Vcore, 38 Degrees Celcius Load

Admittadly the Arctic Silver Three on the JIUHB had been there for about 3 weeks of 24/7 use, and had probably set - and the JIXIB's arctic Silver was freshly applied - but that shouldn't account for such a small temperature difference at a signifigantly lower Vcore.

Also note that my board at the time, an ASUS A7N8X-Deluxe 2.0 couldn't hop a stable FSB above 203 with a multi higher than 11, and 208 with a multi of 11.


From testing two Tbred B DLT3C 1700+ and 1800+ recently:

Both do 2.57 GHz with a TT SFII on a SLK-947U
- 1700+ JIUHB WPMW 0310, 2.57 GHz at 1.95 V
- 1800+ JIXIB MPMW 0332, 2.57 GHz at 1.90 V

This particular 1800+ I have requires 0.05 V less voltage than the 1700+ to do the same 2.57 GHz, very similar to your situation. But I cannot draw too much conclusion based on just two cases from yours and mine. More data points are needed.

As far as temperature, tested the two CPU's using the same motherboard, memory, HSF, ...., within a few days recently. The temperature under Prime95 for system and CPU temperature, using SLK-947U and TT SFII are:

1700+: 2.57 GHz 1.95 V, 18 / 48 C
1800+: 2.57 GHz 1.90 V, 22 / 44.5 C
1800+: 2.57 GHz 1.90 V, 23 / 50 C

What it shows is that the temperature from a CPU under the same voltage and frequency, even within a few days can vary quite a bit, by as much as 5 C (at least from my sampling). If I did not have the second temperature measurement of the 1800+, I might have come to a conclusion that the 1800+ runs cooler supported by a lower Vcore of 0.05 V less.

Given the recent Barton whose week >= 0339 being multiplier locked, the Tbred B 1700+/1800+ seems to be a safer bet, at least for now. Would be interesting to see how are the 1800+'s with week >= 0339.

[Sir Barton]

i need to get another 1800+ to complete my 8RDa+ setup (new BIOS flashrom on order). my chip has hit a brick wall where its at now. seems anything more than 2330MHz is a no go no matter what voltage i hit it with. not complaining though its still a kick arse chip at 2300 and some change.

[BowerR64]

If i ordered one of the 1800s from newegg today what would i probobly get?

would i get somthing that would clock ok? as far as stepping goes.

[sQUiRLy]

If I had to guess, you would get a DLT3C JIXIB 0334 WPMW if you ordered a retail 1800 from Newegg right now.

[sQUiRLy]

What's the last production week going to be for these Tbreds?

[Audioaficionado]

Originally posted by Sir Barton


yep ive backed down my FSB to 200 and tried multis 12 and 12.5 at up to 2.0V and still a no go :( but heck, im happy with an 800MHz OC. 10 hours prime stable at 220x10.5 1.66V :) it will POST at 200x12.5, but windows wont boot.

EDIT: yep, i can push my VDD up to 2.0V with this board.

Try enabling the CPU Interface in your NF7 BIOS. That helps me get more multipliers to work.

[BowerR64]

newegg only has OEM right now. are the OEM chips any good? if i got the 1800 OEM what would i get? if you were to guess

[wannaoc]

The OEM is the same chip you get retail, steppings may differ depending on sales at the reseller though. The difference is there is no warranty and no stock heatsink on OEM chips, thats it as far as I know.

[BowerR64]

This is what i want right?

AMD Athlon XP 1800+ Thoroughbred /266 FSB Processor CPU 1800+/ 1.53GHz -0.13 micro OEM The Ultimate Digital Media Experience in an x86 Platform.Tailored for Microsoft® Windows® XP. Featuring QuantiSpeed Architecture for Rapid Execution of Applications. Tomorrow’s Technology Today.The Power of a Reliable Partner.
Specifications:
CPU: 1.53 GHz
Type: 1800 XP Thoroughbred
Cache: 256K
BUS: 266MHz
Socket A (PGA)
OEM - CPU Only Model#: AXDA1800

$52.

1800+ at newegg (http://www.newegg.com/app/ViewProduct.asp?catalog=343&DEPA=1&submit=property&mfrcode=0&propertycode=&propertycodevalue=4159,3898)

[sQUiRLy]

There's a strong chance that you'll get a DUTC3 if you buy an OEM but then again you may not.

[sQUiRLy]

I bought a retail 1800 about 20 days ago, right before they went out of stock, and got a DLT3C. Most everyone that gets a DL has got it as retail. You'll probably get a revision b either way but depending on your cooling solution you may want to opt for the lower 1.5v DLT3C like I did.

[Audioaficionado]

Originally posted by wannaoc
The OEM is the same chip you get retail, steppings may differ depending on sales at the reseller though. The difference is there is no warranty and no stock heatsink on OEM chips, thats it as far as I know.

Since you're overclocking, you have no warrenty even with the retail. I paid $5 extra and got a year warrenty on my OEM processor just incase it was DOA. The retail HS makes a good paper weight.

[sQUiRLy]

Yea, most places give you 30 days to return an OEM cpu.

I sold my hs/f along with my old Paly.

[felinusz]

The other problem is that AMD has (AFAIK) discontinued the Thoroughbred core processors.

As for the most recent date? Well, we will probably know in a week or two once they start circulating the final batches. I can remember seeing some week 34s though.

[Stinky Bob]

My baby, AXDA1800DLT3C JIXHB 0323WPMW does 2.2 @ 1.6v.

Try any higher, and the voltage req. skyrockets. Temps also climb :mad: , and my PSU (Chieftec 360W) and WC-system has a hard time keeping up.

Watercooled with Magnum 390 block, watertemp. is 30 c.

Bought it here in Norway on aug. 9.

[felinusz]

What do you mean by "voltage req skyrockets"?

If it takes a solid 0.1 mV more to hop 100 Mhz higher than it is performing like it should be ;)

Here is how my old JIUHB 1700+ stepped (I haven't completely "solved" how my JIXIB steps, because the volt mod in my old board prevented me from using Vcore below 1.7, and I haven't got my new board yet):

(These are overclocks obtainable with load temperatures below 50 Degrees Celcius)

1700+:

2200 MHz - 1.625 Vcore (air)
2300 MHz - 1.725 Vcore (air)
2400 MHz - 1.825 Vcore (air)
2500 MHz - 1.9 Vcore (water)

As you can see your processor is very similar - 2200 MHz stable on a 1.6 Vcore isn't shabby. Try going up to ~1.7, and see if you can hit 2300 MHz. Getting to 2400 MHz with a Vcore below ~1.95 is rare, and should be celebrated ;).

[Stinky Bob]

I mean just that.
I can post @2300 1.7v, and @2350 1.775v, anything else it wont have. Brick wall.

Barely posts @2.4 1.95v and temp is 55c. (beep func. set)

[Kekkai]

buying an oem 1800 from Newegg is a huge risk because there are a few people, including myself, who have received

DUT3C

although they were the "B" stepping. I ebayed mine and am looking for places that sell the 1800 retail now, but I cannot find any stores that still carry 1800 in my area.

[Kekkai]

It also depends on how much you want to overclock too. This DUT3C chip did 166mhz x 11.5 at 1.6v but it wouldn't go to 200fsb on an Asus A7n8X rev2. So if you're satisfied with a 400mhz overclock, go for it

[felinusz]

Stinky Bob - Have you tried running with a FSB of 200 even? Try for 200 X 11.5 and see if you get different (better) results - for some reason a really high FSB can be unstable at the same clock a lower FSB is stable at.

[Stinky Bob]

Yeah, I always use low bus while trying to test max core. Settled for 217x10 (for now), nice temps and 1,6v.

[BowerR64]

i just got mine in today from newegg

it was the $52. OEM 1800+ here are the codes

AXDA11800DUTC3 Z335075242017
KIXIB 0338VPMW

what do you think?

[BowerR64]

Im running the same speeds as a 3200+ 2.2gig 1.8 volts, this is totaly nuts! but i love it

195 bus :D

oh ya and its unlocked! this was an awsome deal for $52.

it posts at 2.4 but wont load windows. even at 28c and 1.85 volts

2.2-2.3 is all shes got capin:p

[GBR]

Originally posted by BowerR64
i just got mine in today from newegg

it was the $52. OEM 1800+ here are the codes

AXDA11800DUTC3 Z335075242017
KIXIB 0338VPMW

what do you think?

I received the exact same chip yesterday from Newegg.

Haven't tried it yet, but hope for a decent OC with an SLK 900 and a 92mm fan. 2.2 or 2.3 Ghz and I'll be happy.

Nearly ordered a Barton...lucky me! It was only afterwards that I read the threads that newer Barton 2500 + 's are hard locked at 11.

Looks like a lot more research on CPU's will be needed going forward.

We had it good for a while!

[BowerR64]

thats the same sink i have with an 80mm fan. i got it up to 2.4 but that was only when i sat it out side. I think it got down to about 30F last night. I think 2-2.1 should run pretty stable.


I got it up to about 210fsb but i dont see any performance gain between 210 and 160 seems more stable at the lower fsb

I didnt know till i looked that the 3200s are selling for over $300. :eek:

I see now why AMD is losing money, these chips been clocking that high for what a year or so?

[BowerR64]

I just ordered another one. I think ill give my dad my old 1900+ that bairly does more then 1.7 and put this one in my KT7A

[BowerR64]

Just got the second one in today the codes are

AXDA 1800DUTC3
JIXIB 0339RPMW

Cant wait to see what this one does.

[felinusz]

Wow, that is the most recent 1800+ I have seen to date (Week 39). Strange because we all heard that production was halted weeks ago.

Where did you get it from? newegg?

Tell us how it does! :)

[Audioaficionado]

Make sure you're getting a Tbred B core and not a locked Thorton.

[BowerR64]

Its the same, it does the same speeds as the first one. It can run a little better FSB but runs the same speeds.

I got it from newegg a week after i got my first one.

Only problem is it dont work righ tin my KT7A-Raid, I guess only the XPs will work on tha tolder board. The 1900+ i have runs fine but its an older XP and its locked.

[emericanchaos]

0311 xp1800 JIUHB DLT3C. so far it's doing 10.5*225 = 10.5*225= 2367.87MHz (2951+ PR). motherboard is an Abit NF7-S rev 2.0 in dual channel, CPU interface enabled. memory is still at CAS 2 so i should be able to squeeze some more out of it. going to update my BIOS and change my cooling around as well as some other things.

CLICK: for full story (http://forum.oc-forums.com/vb/showthread.php?s=&threadid=247728)

[StinkBreff]

Do a flash to bios 1.0 w/Sata Fix with FlashMenu and run that sucker at 240 fsb + with cpu interface enabled , all mults.!!!!

[BowerR64]

Originally posted by StinkBreff
Do a flash to bios 1.0 w/Sata Fix with FlashMenu and run that sucker at 240 fsb + with cpu interface enabled , all mults.!!!!

can you link me to that bios file?

[StinkBreff]

NF7d10 with SATA bios v4.2.27 (By Equito) (http://www.tnplanet.com/nf7/bios/2/snf7d_10.zip)
Fixes 10 and 10.5 multiplier in most cases..
You'll also need Abit FlashMenu 1.30 (http://www.logan.eclipse.co.uk/Downloads/FlashMenu130.exe) to flash the bios backwards thru windows .. After bios flashing is done , shut down , pull power , clear bios for 20-30-60 seconds .. Move jumper back .. Power up and rock the block til your mobo locks !! Lemme know how it goes !!

1. Choose to flash from FILE. Note: it is recommended to download the bios file to your hard drive and place the “dot bin” file in the FlashMenu folder.

2. Browse to the bios “dot bin” you want to flash.

3. Choose Open and the flash will begin.

4. Once finished it will ask to reboot. DON'T just yet, choose NO and repeat the flash again, YES that's correct do the flash twice so that it takes better. It helps with instabilities.

5. Now reboot when it asks if you want to Reboot this time.

6. If it fails to fully boot where you can enter the CMOS for setting, just shut down completely with the power button. Hold it in for ~5 seconds or until the computer shuts down fully. Turn off the P/supply switch on the rear @ P/supply or remove the power cord from the wall outlet or the P/supply. While the power is completely removed from the computer you should push the power button in on the front of the computer and usually the fans or a light on the front of computer will come on momentarily to discharge the P/Supply.

7. Now re-establish power to the computer AND while holding down the INSERT key on the keyboard you press the power button on the front of the case and it should boot up with defaults set and allow you to enter the bios at post using the DEL key and set your cmos however you wish (for help see bios settings)

NOTE: DO NOT choose to flash from Internet as there is too much chance of it Fcuking up.

NOTE: DO NOT OC when attempting to flash a bios. If you’re not sure about flashing then, it is time to learn the correct procedures, as it is a near necessity to flash the bios at one time or the other when seeking the ultimate performance available in the NF7/S motherboard series.

NOTE: The best way to flash back to an older bios revision is with FlashMenu so use it. It can be done with DOS and a Floppy but FlashMenu is easier.

NOTE: If flashing the bios rom a lot, or pushing OC to the edge, then it might be advisable to purchase a backup bios rom chip, or if you have the monies purchase a BIOS Saviour.....

[felinusz]

Man, thanks for the file and great info - I will try it out tonight and tell you how it goes. Right now I'm stuck at 211 with CPU Interface, and APIC enabled.

My JIXIB is currently at 211 X 12 - 1.85 Vcore, 24 hours prime95 stable.

[Minjin]

Not to go off topic too much, but my NF7-S with bios 19 (newest one) has no problems with any multiplier. I don't understand why people are talking about 10 and 10.5 problems. Works perfect for me.

...and yes, CPU interface is enabled.

Mark

[BowerR64]

That modified bios file is for NF7-S rev 2 right?

before i hose my board i wana make sure its the right one.

[StinkBreff]

Yup - Nf7-s v2.0 only :attn:

[felinusz]

*OFF-TOPIC*
Well I tried it and it made no difference, I'm still stuck at 211. Haven't done the L12 mod, or the VDD Volt mod yet though so.... I'm sticking with this BIOS for the time being.

[hitechjb1]

Originally posted by Minjin
Not to go off topic too much, but my NF7-S with bios 19 (newest one) has no problems with any multiplier. I don't understand why people are talking about 10 and 10.5 problems. Works perfect for me.

...and yes, CPU interface is enabled.

Mark

From my experience overclocking two NF7-S rev 2.0, which has the best FSB potential, but some are very sensitive and require all kinds of trick to get it to work at high FSB.

- One failed mysteriously at multiplier 9, 9.5. 10 (hang, crash), could only do 215 MHz FSB. But with the right bios (version 10 only which is not recommended due to bios corruption), FSB_sense 166 mod, CPU interface enable, it can do 230 FSB with memory efficiency at around 95%.

- Another one can do 225 MHz without any trick and mod and bios change. But nothing can help it to go higher (limited by memroy controller in NB). Even with this one, there are problems with multiplier <= 10, memory efficiency suffers by 5-7%.
I only need mulitplier 11 or 11.5, since my CPU's can run 225 x 11 or 225 x 11.5, so multiplier 10 is not a problem for me.

Even you think your bios can handle multiplier 10 or under, I think better check the memory efficiency. It should be around 95% for this motherboard w/ CPU interface enable. If you get memory efficiency of only 89%, it is equivalent to losing 12-15 MHz FSB.


For details about the testing, effect of various bios version, bios setting, FSB_sense mod, memory efficiency, ...., click this link:
Summary on overclocking the NF7-S rev 2.0 (with Tbred B 1700+ DLT3C and ...) (http://forum.oc-forums.com/vb/showthread.php?s=&postid=2154286#post2154286) (page 15)

[BowerR64]

Whats the difference between 150 fsb and 200?

I ran a bench on both settings and i dont see any difference with my machine.

Running the same speed mhz wise they werent any different in the benchmarks :(

[Audioaficionado]

You should see a better memory score as there's a higher RAM/NB/CPU bandwidth with 200fsb.

[deeppow]

Does anyone know of a good web retailer from which to buy the 1800DLT3C? Only one I can find is eXcaliberPC which has the JIUHB:0311MPMW. A number sell 1800s but upon visiting their site you can't really tell what stepping or week.

Thanks for any help. :cool:

[StinkBreff]

I've got few I could let go .. Pm me for details as I cant post in sales forum as of yet .. 100 posts ... jeesh .. Let me in admin/mods !!
Jiuhb looks like pic , different week though

[felinusz]

Another good source (although mine are out) is local computer stores/shcaks. They often have lots of processors to choose from, so you can grab a prefered stepping. You might have to travle a bit, and visit a lot of stores, but you will probably be able to find a T-Bred in a recent week that way too.

Buying from another member is the best way thoguh ;)

[StinkBreff]

PM me and I will tell you deep .. I dont want everyone buying them all cuz I wanna buy another soon ..

[clawhammer]

Ok, heres mine :
1800+ DLT3C NIUHB 0335 UPMW @ 238 x 10.5, Vcore @ 1.8
Epox 8RDA+ :D
Prime & 3dmark stable !

[GWN]

Just picked up a 1800+ JIUHB DLT3C 0329

Just letting it run overnight at 2.01 Ghz @1.5 volts to let it settle in. So far so good. I will see how far I can push it. I was quite surprised to see that it runs at 1.5v at stock. I wasn't sure if this was correct. Is it? I thought they ran at 1.6-1.65v!

[StinkBreff]

1.5v is right ! My jixib and jixhb even default to 1.47 .. You will problaly get about 2.1 - 2.2 tops at stock voltages .. Possibly up to 2500 mhz at 1.85 .. Depends how how well you have it cooled , and how stable your motherboard is also ..
Then again you have water - maybe 3 ghz will be your limit ..

[GWN]

Thanks for the heads up on voltage. I will play with it tonight and post some results. Looking forward to it. My temps don't seem to move too much no matter how many volts I pump through my chips so hopefully I can really crank it up.

[BowerR64]

This is a cool chip, one of my favorites. The 2.4c and the 1800+ are 2 fun chips we have out now.

2.4c= $175.
1800+ = $50.... :eek:

[GWN]

ok now running at 2.34 Ghz ( 166x14) @ 1.675 volts. Going to let that run for a while and then see how much more I can go. It's starting to get toasty!

[felinusz]

All DLT3C thoroughbreds - I.E. these ones - stock out at a Vcore of 1.5 - your BIOS and MBM5 may well read it as 1.47, but that is simply your NF7-S undervolting your processor (I know that mine does).

clawhammer: That is a wicked awesome fast FSB overclock. Hopefully I can net similar results after a L12 and VDD voltage mod :/ :).

StinkBreff - I highly doubt that any non-phase cooled AMD processor will ever hit 3 GHz, let alone stably :D ;).

GWN - try for 2400 Mhz at a 1.7 vcore and see how stability is, you should hopefully be able to hit 2500 MHz with a Vcore range of ~1.75-1.9, and 2400 Mhz around ~1.6-1.75.