venice on ice 2850mhz 1.38v -40°C

Quote:

---

Originally Posted by pc ice

air and water result's from the 3500 will be posted in just a few hr's.

---

Great :)

gogo :p:

It's very nice
I think on air those 3000/3200 can climb 2.7ghz easily.

sweet avatar for wonder bra

Quote:

---

Originally Posted by s7e9h3n

Th other 10% of us are wondering where are the XTREME screenies of this 3500+ :D

---

I think that air and water are more important cause there is no real use for running these under extremly low temps (-70+) cause u cant use them 24/7.

Most people run on air / water.

:)well i just have to by a 3500+ venice

Quote:

---

Originally Posted by reject

wat is sSOI?
ive only heard of SOI (silicon on insulator) SS (strained silicon) and SSDOI (a combination)
winchester doesnt have SS so that theory doesnt work :)

---

i talked to amd on cebit, and yes, all their 90nm cpus are built on soi and use SS. at least thats what they told me.

Quote:

---

Originally Posted by matt9669

The FX-55 is and always has been strained silicon-on-insulator or sSOI, AMD publicly declared that at launch :up:

As for the cold bug - there are a number of theories but I personally believe it has something to do with the thermal diode reporting incorrectly, not inherent limits of the chip - AMD changed the diode with Winchester, they changed it again with Venice AFAIK and since we haven't seen new BIOSes for the chips yet (they haven't been officially launched, go figure) it's quite possible this issue could be resolved for all Venice core CPUs.

I remain optimistic! :rehab:

---

hmmm good idear matt! maybe the new chips need an updated bios as well to get rid of the cold bug completely... could be possible as well...

but from what i heard the temp probe hasnt changed in venice, it only has been changed in winchester and rev E4 aka san diego, wich is why they both needed a new bios to boot while venice doesnt need an updated bios.

Quote:

---

Originally Posted by R.Rabbit

yeah i didn't think about that... hmmm, well its gotta be the diode, i wouldn't be surprised if this cold bug is an anti overclocking issue made by AMD. it would make sense becuase they designate the FX line as their enthusiest proc and conveniently(spell?) it has no cold bug... similarily as we all know its mult is unlocked
but if this where true if then logically there would be a specific temp in which the procs would stop working and all would have the same bug, not just some :shrug:

---

lol, funny theory, but no, i really really dont think amd tried to limit ocers by making their chips unstable or not bootable at low temps. if they did then they did a really bad job since many chips seem to still boot at -40°C or even colder...

and i dont think amd would change their cpus just because 0.01% of their cpus get supercooled and overclocked, wich helps them marketing and selling their cpus...

Quote:

---

Originally Posted by bugeyes

Those 256MB sticks might have had chips on both sides of the memory module
but are in fact single "banked" alot of people get confused between double banked and double sided DDR SDRAM...

To test the memory controller they will need to use 4x 512MB ram modules, these will be double banked.

---

there are also DS SB 512mb sticks, but very few.
pc ice said he will use 4 512mb sticks of tccd and see if they run at stock with 1t and how high he can get them :)

hey MsB, welcome to XtremeSystems :toast:

Quote:

---

Originally Posted by Ackbar

Kinda off topic... but I just want to say that Saaya has been doing a good job of posting this topic to other forums! I'm sure a lot of people in other forums that don't post on other forums are VERY interested in the happenings of the Venice and don't always know where to get the best info. :toast:

---

thx :)

many news sites i sent a link to the venice preview didnt want to post about it because they or their partner sites will have venice reviews ready soon and fear that less people will read the review if they already saw a preview :rolleyes:
so i thougt i will just bring the news to the people directly then :D

Quote:

---

Originally Posted by Hallowed

This chip has been tossed around for what... 4 days and noone has bothered to test 4x512 1T yet?

Its the burning question we all have. :stick: :(

---

pc ice hasnt picked up the chip from opp yet, and opp can get it to boot under his cascade...

welcome to XtremeSystems Gsus_Freak :toast:
and im sure amd likes ocers, if they wouldnt then they would not sell a chip like the fx with all multipliers unlocked, and they would have locked all multipliers on all a64s and not just the higher ones ;)

Quote:

---

Originally Posted by HaLDoL

4x256 BH5 was tested and it wouldn't boot

---

this sounds more like a bios problem to me though, since bh5 only comes in 32mb/ic sizes wich means 4 256mb sticks bh5 are pretty much like 2 512mb sticks bh5... of course there are other differences, but whether its 4 256mb sticks or 2 512mb sticks, its still 32 memory chips to address either way... so kinda weird the one works but the other doesnt?

Quote:

---

Originally Posted by Rolle2k

s7e9h3n: Did you test how high the FSB did go with 4x512 and 1T timing? cus im going to order 4x512 TwinMOS Speed Premium & A64 with Venice core soon, and would like to know.

---

please read some posts before you ask something, in every page of this thread this question has been asked at least once and at least been answered once as well :D

amd wouldnt block the procesors for -100°C people that have that equipment is like 0.000001% of the market and limting the procs. because of them would be expensive

that would be bad for marketing couse most of us come here and see pc ice, fugger, macci, opp, opb, etc... and go and try to buy the procesors and all we can buy to reach the same level of oc (thats insane but hope is the last thing you loose)

p.d. sorry for mi english it sucks :(

He said they'd be posted in about an hour... it's been nearly 3. :(
I guess it's worth it though, I just hope he's not running into problem's right now. :eek:

Quote:

---

Originally Posted by Orthogonal

He said they'd be posted in about an hour... it's been nearly 3. :(
I guess it's worth it though, I just hope he's not running into problem's right now. :eek:

---

maybe bad traffic on the streets between opps and pc ice's place :D
just subscribe to this thread and check it out as soon as you get a notification of 3 new posts in this thread in a row, than you can be sure that something new has been posted! :D :lol:

And he said a few hours so 2 should be considered a minimum, patience ;)

hehe can't wait...

Quote:

---

Originally Posted by LowRun

And he said a few hours so 2 should be considered a minimum, patience ;)

---

2 is a couple. A few would be considered at least 3 hours. :D

I have been communicating with an AMD Technician over the past several weeks over various issues, and hopefully this will clear the air in regards to filling all 4 DIMMS and running 1T CMD rate.

I asked this:

Will the Venice and SanDiego CPU's resolve this
issue.........meaning will these cores allow all 4 DIMMS filled with
double-sided ram on s939 boards to default to DDR400 and 1T CMD Rate
instead of DDR333 and 2T like the Winchester/NewCastle/Clawhammers
currently do?

His reply was this:

As you have perceived, the Venice and San Diego cores will alleviate
this problem. AMD has revised and improved the memory controller
within the chip to eliminate the issue.

Hope this helps. If you have any other questions, please feel free to
contact me.

Regards,

Quit stressing..................it will work. :D

Thanks Lidocaine, damn good news, my 4x512 sticks should be here in 2 days so i'm a happy camper :D

Thanks for the news Lidocaine :banana:

thx for the infos :)

hmmm in case the 4 512mb sticks DONT work with 1t on pc ice' rig, then this means either your amd guy misunderstood you and meant they will default to ddr400 2t and not ddr333 2t like they do know, OR the boards need a bios update to not force 4 512mb sticks to 333 or 400 2t, wich makes sence as well....

if we're lucky it will work with the current bios :D

is it me or has it been almost 6 hours since PC ICE's post :stick: :rolleyes:

I CANT WAIT FOR THEM TO SHIP ON THE 4TH :p:

there is one thing that bothers me tho everyone says the cold bug is the chips fault me personally think it is a mb issue that a simple bios flash will fix like how every one has to flash there bios to be able to use the venice.

i just dont see why AMD would not work on this issue more other then pass it down to a nether core well hopefully they will get it all sorted by the time i get mine because i wouldnt mind putting it under a ps unit.

Quote:

---

Originally Posted by Fzero

I CANT WAIT FOR THEM TO SHIP ON THE 4TH :p:

there is one thing that bothers me tho everyone says the cold bug is the chips fault me personally think it is a mb issue that a simple bios flash will fix like how every one has to flash there bios to be able to use the venice.

i just do see why AMD would not work on this issue more other then pass it down to a nether core well hopefully they will get it all sorted by the time i get mine because i wouldnt mind putting it under a ps unit.

---

It's because there essentially is no demand for a cpu that functions sub-zero. Only a small (I mean real small) percentage of users put cpu's under phase-change. ;)

About the temp issue... everyone keeps saying only the E4 (San Diego) CPUs had an updated thermal thingy. The E3 (Venice) remains the same (a la Winchester cold bug).

Therefore, it is my hypothesis that AMD redid the temp problem for their E4 90nm stuff so that people will be able to phase change cool FX-57s for some bad ass overclocking. We poor losers that have to get Venice CPUs (E3) will be relegated to -90 C or higher temps. :) (I use -90 C because that is when Opp said the Venice stopped working under his cascade.)

Remember, AMD is far more concerned about enthusiasts than some other CPU companies. Being that the FX-57 is an enthusiast-oriented CPU, it is to AMD's benefit not to thermally cripple it.

from what i heard it didnt stop working at -90°C, it didnt work at -90°C!


so far we only got confirmation of it working at -40°C and not woking at-90°C afaik, somehwhere between that it stops working, but who knows where...

and yeah mcnbs, about the temp probe rework and the fx57 thats exactly what i think, they re-did the temp probe so their enthusiast cpu is ready to reach some nice numbers.

I have come to the conclusion that PC ICE got kidnnapped by aliens theyre probably probing him now... :stick:

Quote:

---

Originally Posted by winston64

I have come to the conclusion that PC ICE got kidnnapped by aliens theyre probably probing him now... :stick:

---

Oh. I thought he got dragged into the jungle by drunk monkeys...

Quote:

---

Originally Posted by saaya, referring to testing 4x256MB dimms

this sounds more like a bios problem to me though, since bh5 only comes in 32mb/ic sizes wich means 4 256mb sticks bh5 are pretty much like 2 512mb sticks bh5... of course there are other differences, but whether its 4 256mb sticks or 2 512mb sticks, its still 32 memory chips to address either way... so kinda weird the one works but the other doesnt?

---

But wouldn't signal quality decrease when it is split between 4 banks instead of 2 banks? If what you say is true, then we would be able to oc 4x256 to simliar frequencies reached with 2x512. Who knows, maybe we can and nobody has tested...but I'm pretty dern sure it would take extra voltage. Nonetheless, it could easily be a bios problem.