• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

A64 Mobile 4000

Permabanned
Joined
31 May 2005
Posts
1,628
Location
South Yorkshire
i'm sure one or two people around here have been expecting this post (mentioning no names - you know who you are ;))

Yesterday I received the above processor. Having read as much as possible about the thing and support for it, I found out that Epox don't officially support any Turion or Newark E5 rev. cores for their desktop boards, and neither will they write in any support for it in future BIOS releases. My board is the 8N-PAJ PCIe.

However, on the back of reading a post on another board detailing a way to get these CPU's working in an Epox board, I went ahead and purchased the processor anyway.

Swapping out my 3700 clawhammer for the Newark caused the board not to POST, and the onboard error LED's showed a CPU error - which I expected.

The method to get the Newark to work is to disable cool n quiet and manually adjust the multiplier to the lowest (4x) whilst having the Claw in, then pop the Newark in. Then boot into windows at 800mhz - use whatever utility to increase the voltage and multi up to stock - which then enables you to adjust the multi and voltage in the BIOS, to ramp the chip up to stock - hey presto one working Mobile 4000 Newark ripe for overclocking.

Fair enough - everything was going rather swimmingly until I tried to adjust the voltage through windows - every time I tried it, the system hung and needed resetting. The multiplier could be upped to 12x, but it would hang at 13x. I used CPUMSR and CBID to attempt this under windows.

I also tried changing these settings through the BIOS, which caused some very strange results. No matter how many times I tried to change the voltage, it always reset itself back to 1.1 (stock is 1.35) and occasionally just wouldn't POST at all. The multiplier would change, stepping up in 1x increments. at 7x and 8x the BIOS showed stock voltage and the corresponding multiplier, yet as soon as i went past 8x it reverted back to 1.1v - then the CMOS needed clearing as soon as i went to 13x as the system refused to do anything at all.

The CPU reports itself correctly in CPU-Z, Newark mobile core, SSE3, 90nm etc. yet the only thing that seems to be holding me back is the system hanging as soon as I try to up the voltage in windows.

Now I know i'm going to get the 'why did you buy it if you know your board didn't support it' replies, but i've kinda got half way there by making the thing POST and being able to do certain things like up the multiplier in the BIOS and I've seen it report the voltage correctly (as well as correctly retaining my RAM timings which doesn't happen when it POST's at 1.1v).

So, anyone have any ideas? :D
 
Looks like the DFI nF4X Infinity S754 board will be the best bet, from reading around.

However, none available in the UK as far as I can see :)
 
Hi Andy,
Don't tell me that by putting that smilie in you are now going to try to get that DFI from the states as well. You must be a glutton for punishment. Is that an Sli board too?

I must say that after reading that article on how to get it to work in our mobo i was very tempted, but i am now glad i waited to see how much hassle you had. I think i will stick with my 3700+ Claw. Did you use the same piece of software mentioned in the other forum?
 
DFI don't officially support the mobile processors either but they work.

I'd say check your headsink is making good contact with the processor because I had issues with the system cutting out and resetting the multi and voltage settings in the bios.

Also flash to the newest bios, etc, blah blah blah!

Sorry for stating the obvious...

Edit: Remember you have to mod the bit around the socket. *Sorry again*
 
Yeah, you take off the top part and rub it over sandpaper to take a couple of milimetres off the legs.

I just took mine off though and bolted on a storm water block but any bolt on heatsink will do the trick.

To see if that's the case turn the computer on it's side so the weight of the heasink is pressing onto the cpu then set the bios.
 
There you go Andy unless you have tried it already. Let us know if it works.

Have you still got yours Meths? Last time we heard you were waiting for your watercooling bits before you went any further.
 
pastymuncher said:
Have you still got yours Meths? Last time we heard you were waiting for your watercooling bits before you went any further.
I've still got it but tbh I've been running it at stock speed most of the time.

At one stage I had 1.7V going through it :eek: but it's maxing out around 3.2ghz which is totally unstable.

Andyp getting his is spurring me on to get a decent overclock on it so I'll start a new thread with some benches shortly.
 
I'm using a DFI UT nF3 250Gb board with a mobile CPU, see below :)

When I first got the chip, it POSTed at 800MHz and 1.1V, but the DFI board was happy with me going into the BIOS, setting it back to what it should be, and restarting :)

Had a quick go with it, using an XP-90 and YS-Tech fan, managed 220FSB before the memory crapped out (was value ram from an overpriced high street retailer) and then gave up and ran it at stock.

I should have another go though, now I have some good OCZ stuff :D maybe I will tonight. I know the room temps in my bedroom are bad though, it's always too hot in here for some reason :(
 
Well the plot thickens.... surprisingly enough :)

I tried one more time to get the 4000 working, and it wouldn't even POST. The mobo onboard debug LEDS showed an 'FF' error which is a CMOS problem according to the manual. So I pretty much gave up on it - and popped the 3700 back in. Cue one screwed BIOS - I reset the CMOS, took the battery out and all that shenanigans. The board then POSTed fine, but it seems stuck at the factory default settings, CPU overvolted 0.1v, RAM at 3-3-3-7 2T - any attempt at all to change this hangs the system and it defaults back. Then Windows wouldn't load.... at all.

Naturally frustrated (the Newark nearly ended up as a rather fetching key fob) I've just finished a complete reinstall of XP and seem to be stuck with a rather crippled system. It works fine and seems stable enough, but the BIOS defaults are stuck it seems.

I've been in touch with Epox who have reiterated what I've already read elsewhere - they won't, and they don't intend to support any of the rev. E5 CPU's (although E6 support was written into an updated BIOS last year). So to hell with them. They claim to 'be the best for compatibility and support' yet they won't even write in (a process which I gather for someone in the know, isn't really that hard) support for what is essentially the ultimate performance component for Socket 754. I've spent quite a number of hours looking around for unofficial BIOS's for the Epox and there's none to be had anywhere.

So at the moment I'm pretty stuck. I loved my DFI NF3 250GB board when I had it, and up until now I've been pleasantly surprised by this Epox board. The Biostar board is Micro-ATX so I don't think I'll be going down that route. Given my previous jollies with DFI I think, although rather inconvenient, it seems like the best bet.

Although, it's certainly not all plain sailing from there. The great people over at DFI street (if I'm not allowed to say that, shout up and I'll remove the mention) are as helpful as always and one of them had this to say...

I feel your pain with the Epox. I dumped my 8KDA3J because of a lack of Newark support and, of course, because it didn't have PCI-E.

Just FYI, here's what a poster at *** named ****** said about Newark and the NF4X:

"Another quibble is the bios doesn't fully support my Newark. It won't restart....I have to power down each time."

*deep breath* Soooooo...... :D
 
On the other hand - upon doing just a few minutes reading about these here Biostar T-Force 6100 boards, there's some huge overclocks and virtually flawless support going on with the Newarks. And I can pick one up for around 40 quid in the UK.

Stupid question coming up - will the Micro-ATX form factor of this board be a problem with my Antec Antec PlusView 1000AMG case? As much as I want to get this whole thing up and running, I don't want to be using sellotape, string and blu-tack to hold it all together :D

*edit - and I have a rather huge Leadtek 256MB 7800GTX to squeeze in there as well. The bottom edge of the card rests on the Epox's (used) IDE connector and causes the PCIe socket-locking connector thingymabob (i love technical jargon :D) not to click into place properly. The card is seated well enough and doesn't waggle about or come out of the socket, but it's just not 100% secure. Wonder if there'll be any problems there?
 
Last edited:
Ah it's fine, all this is part of the fun for me.

The thrill of the chase and all that.... :D
 
Back
Top Bottom