venice on ice 2850mhz 1.38v -40°C

Quote:

---

Originally Posted by Orthogonal

I've asked that question 2 times already in this thread and noone seems to know :mad:, or atleast won't answer. I'd like to know as well

---

IIRC, didn't he say he already had 3/10? Someone gave the idea of going back in the BIOS updates to see at what point they stop working and he thought it was a good idea...so maybe that's to come?

Quote:

---

Originally Posted by Ackbar

Two CPUs and you're ready to make a generalization about the whole lot of CPUs not having a cold bug? I have no information one way or the other, but let's see what happens when these chips come out. I wouldn't be surprised if these CPUs were not typical of the lot of chips. Until I hear from several users, I'm going to hold out judgement about how well these things do in general.

---

i heard of only very few winchester chips that dont have a low temp bug, and actually they still have a low temp bug, they just boot at low temps, but not really low temps. from what ive heard they only go down to -30°C max and dont work at lower temps, while now we have a pair of cpus that works flawlessly at low temps down to -40°C so i think its very promising... :)

Quote:

---

Originally Posted by saaya

i heard of only very few winchester chips that dont have a low temp bug

---

Looks like venice is the same, OPP said it wouldn't play nice in -90s :eek:

Is -40c the coldest you can take these?

Quote:

---

Originally Posted by saaya

i heard of only very few winchester chips that dont have a low temp bug, and actually they still have a low temp bug, they just boot at low temps, but not really low temps. from what ive heard they only go down to -30°C max and dont work at lower temps, while now we have a pair of cpus that works flawlessly at low temps down to -40°C so i think its very promising... :)

---

I have had 3 or 4 winnies. All bootet up and ran just fine at low temps. At boot -50c and load -44c on evap. All from different weeks. (week42, 45, 47 and 50)

I still thinks its a mobo issue cause if I start it up and then start the cooler it locks up when it hits 0c.

Quote:

---

Originally Posted by isp

Looks like venice is the same, OPP said it wouldn't play nice in -90s :eek:

Is -40c the coldest you can take these?

---

its single stage! lol, yes -40°C is the lowest his single stage unti can pull i think, but hes getting his cascade ready (-100°C load temps :D)

Quote:

---

Originally Posted by Formann

I have had 3 or 4 winnies. All bootet up and ran just fine at low temps. At boot -50c and load -44c on evap. All from different weeks. (week42, 45, 47 and 50)

I still thinks its a mobo issue cause if I start it up and then start the cooler it locks up when it hits 0c.

---

really? now thats interesting... you tried them on the dfi board, right?

Quote:

---

Originally Posted by Formann

I have had 3 or 4 winnies. All bootet up and ran just fine at low temps. At boot -50c and load -44c on evap. All from different weeks. (week42, 45, 47 and 50)

I still thinks its a mobo issue cause if I start it up and then start the cooler it locks up when it hits 0c.

---

I tested 3 different winnies, 1 3200 and 2 3000 on gigabyte k8ns (plain, not ultra) and DFI Ultra - d and all of them booted @ ~ -30c @ evap... plus al overclocked far higher than with air.... 3200 and 1 3000 cbbid plus 1 3000 cbbhd. I also think it might be something on the mobos (not all mobos off certain brand / model but some).

hmmm very interesting then...

Quote:

---

Originally Posted by IvanAndreevich

You two are blind. 3800+ was ran on Neo2 with 1.37mod bios which is very old. Bios is only required so that the chip is detected properly but it will WORK FINE.

---

I realized that, but that has nothing to do with this thread :stick:. I was asking about his 3500+ on the DFI Board. I've heard things on DFI-Street that the 3/10's are known to work on the new chips but they weren't sure about older BIOS's, so I wanted to know if he was perhaps using an earlier version

Quote:

---

Originally Posted by saaya

hmmm very interesting then...

---

this week i have 2 more 3000 and a different gigabyte mobbo to try (thing is they all go to customers, and if something goes wrong, i should pay for them (the chips, that's it....) but... hell all goes in name of overclocking sience
:D

Superb, I can't wait to get me one of those :toast:

bad news! opp cant get it to work under his cascade, seems there is still a low temp bug :S :(

now lets hope E4 changes that, so far i only heard that E4 had changes made in the temp probe , thats why it needs a bios update to work on almost all boards.
i really hope the low temp problems dont have to do with the manufacturing process! :(

Quote:

---

Originally Posted by saaya

bad news! opp cant get it to work under his cascade, seems there is still a low temp bug :S :(

now lets hope E4 changes that, so far i only heard that E4 had changes made in the temp probe , thats why it needs a bios update to work on almost all boards.
i really hope the low temp problems dont have to do with the manufacturing process! :(

---

My guess is that the "low temp bug" has to deal with the size of the core. I've been thinking about this a little, and I have a theory about why 90nm and smaller may not do well with the cold, but it's rather unsubstaniated.

well if it is about the diode i know of a fix... the two pins that read the diode temp on the mobo are on the outside edge of the cpu. just bend them outwards.. if you want to read the temps you can make some hardware to read the temps.. or if you dont care what the cpu says its temp is you can just plain cut the pins off.. no harm done.

Quote:

---

Originally Posted by Ackbar

My guess is that the "low temp bug" has to deal with the size of the core. I've been thinking about this a little, and I have a theory about why 90nm and smaller may not do well with the cold, but it's rather unsubstaniated.

---

and why prescott go crazy with LN2 at -190° ??:stick:

i think is amd that is doing/doing wrong on something...

Quote:

---

Originally Posted by Malachia

and why prescott go crazy with LN2 at -190° ??:stick:

i think is amd that is doing/doing wrong on something...

---

Well, like I said, it's all just an idea that I have. But I'm guessing that it's partly to do with the low power envelope of the AMD 90nm, not necessarily innately to do with the core size, but it's a consequence of it obviously.

Quote:

---

Originally Posted by saaya

bad news! opp cant get it to work under his cascade, seems there is still a low temp bug :S :(

---

That's what I was sayin earlier...damn cold bug :mad: :(

We need to find out what the coldest they can go before they start flaking out

this could have something to do with sSOI itself, i think i read a the begining when sSOI was first being theorized(atleast by us) that it may have a problem with colder temps and would become unstable, but this problem doesn't exists with prescott.. maybe its just AMD's sSOI..?

Quote:

---

Originally Posted by MaxxxRacer

well if it is about the diode i know of a fix... the two pins that read the diode temp on the mobo are on the outside edge of the cpu. just bend them outwards.. if you want to read the temps you can make some hardware to read the temps.. or if you dont care what the cpu says its temp is you can just plain cut the pins off.. no harm done.

---

maybe some one with a winchester could try this? the solution would be the same if this is the cold bug problem, so if someone with a winnie tried this instead of venice(much to valuable at this point) that would be great

Somehow I have a feeling that "Venice" core is not gonna bring more hope to cascade users than "Newcastle" and "Winnie" do now. We have seen some Winnies doing benches at 3.0-3.1GHz on phase change, though. Venice might be good for moderate-low temps but looks like too much cold renders them useless. Unless, there IS something that can be done to make it work. Let's wait for more information from OPP...

On the side note from what we see here now... I hope I'm REALLY wrong but I wouldn't be surprised when FX-57 will not match speeds FX-55 now does under cascade... Hopefully, I'm seriously OFF!

EDIT: Not that Winchester or Necastle were ever chips of choice for cascade users...

i am quite interested to know the difference between air and phase change...

Quote:

---

Originally Posted by Malachia

and why prescott go crazy with LN2 at -190° ??:stick:

i think is amd that is doing/doing wrong on something...

---

intel 90nm =! amd 90nm ;)

Quote:

---

Originally Posted by isp

That's what I was sayin earlier...damn cold bug :mad: :(

We need to find out what the coldest they can go before they start flaking out

---

it varies from chip to chip it seems...

Quote:

---

Originally Posted by R.Rabbit

this could have something to do with sSOI itself, i think i read a the begining when sSOI was first being theorized(atleast by us) that it may have a problem with colder temps and would become unstable, but this problem doesn't exists with prescott.. maybe its just AMD's sSOI..?

---

never heard that they became unstable, they just didnt clock higher...
2ghz air and 2.2ghz ln2 was something common for the first a64 chips :D

nothing from opp yet... hope he posts some infos soon :)

Quote:

---

Originally Posted by saaya

really? now thats interesting... you tried them on the dfi board, right?

---

Nope. Haven´t tried any in my DFI yet. Just in the Neo2 0411 Korea.

No problems what so ever. They all overclocked average. 2500-2600 on air/water, and 2900-2950 with the CPU cooler (-44c load)

Anyone got a Winnie that cant handle the cold? Give me a fair price, and Ill buy it just to prove my self wrong :)

Would be nice to get it confirmed. That its/its not just mobo issues i mean.

Quote:

---

Originally Posted by Ackbar

Well, like I said, it's all just an idea that I have. But I'm guessing that it's partly to do with the low power envelope of the AMD 90nm, not necessarily innately to do with the core size, but it's a consequence of it obviously.

---

Maybe maybe not.
Problem more likely is diode location inside the chip or how it is programmed. Different board will impact bootability (prime example is MSI vs Asus A8V or something 939). I've had (dis)pleasure of finding out Asus booting maybe 10-20c colder temp than MSI on same chips. Also, each winnie has different behavior on same mob. Some will boot till -20c, some till -30c, some till -50c, some will last till -60c (all evap temp) and few of them won't do anything below 0 (had 1 that actually I couldn't really find out how low as my single phase unit didn't go below -70c on evap). Now all these temps are about 5-10c off from actual evap temp is my guess even for heavily tuned Mach II unit with baker evap but there it is. Maybe venice plays nicer but who knows. There were few FX-55 that couldn't handle cascade as well. Only time will tell.

this doesnt look good for my future chilly1 unit, i might have to pick up a 3500+ clawhammer or something

Quote:

---

Originally Posted by agenda2005

WHY IS EVERYONE WITH THIS CHIP HIDDDING AWAY OF SHYING AWAY FROM RUNNING 4 MEMORY DIMMS AT 1T CMD AND DDR400.
All other benchmarks shown here are not new or innovative in anyway. No offence intended.
PLEASE RUN DOUBLE SIDED 4 X512MB RAM at DDR400 and 1T CMD SHOWING ALL DIMMS IN SANDRA SHOT SINCE CPU-Z IS BUGGY. HOW LOUD DO YOU GUYS WANT US TO SOUND ABOUT THIS ISSUE?

---

Hey bro,
I don't know if you've read through my thread about the Venice 3800+, but I CLEARLY stated that I DID try to boot 4x256 bh5 and it DID NOT WORK. I then tried only 2x256 and could boot then. BTW, the 256 sticks were all double sided. Hope that settles that.... :rolleyes: