Yeah i agree with you fully! unbeatable Power/Heat(ultimately noise) ratio!!! definately.TomMM wrote:Perfect! Forget all this overclocking stuff - this is a great SPCR CPU.
Mobile Barton 2500 - owners and wannabe owners...
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Smokeey I think I read somewhere that you can get bartons to run in the k7s5a but something was missing. Should be same for mobile. You could get a 2200XP-M which uses a T-Bred core and it would run fine.
Also, Im using an Abit Nf7-S v2 and at bootup and in SiSandra it says 'unknown CPU type' but Aida32 sees it no problem.
Everyone else get the same thing?
Also, Im using an Abit Nf7-S v2 and at bootup and in SiSandra it says 'unknown CPU type' but Aida32 sees it no problem.
Everyone else get the same thing?
Thanks for the reply PhilgB. I've done some research on the ocworkbench forums and people on there have managed to successfully install the XP-M chips in this board. My initial doubt stemmed from the fact that the K7S5A does not accept desktop Bartons due to the 166/333Mhz bus frequency (the K7S5A can handle upto 133/266Mhz max with the 'official' bios).
However as the mobile Bartons run at 133/266 frequency they should work without issue! The only niggle is the is the voltage - from what I've read it appears as though the K7S5A cannot supply a voltage as low as 1.45v - which isn't a problem if you are overclocking - but in my case I just want to run at stock speed and have a quiet system so that may be the limiting factor
I'm still researching though so there may be hope yet! Makes for some very interesting reading though
EDIT: My mistake - is does support 1.45v Its just that it doesn't allow modification once the default CPU voltage is detected - no issue for me as I dont plan to overclock...woohoo!
However as the mobile Bartons run at 133/266 frequency they should work without issue! The only niggle is the is the voltage - from what I've read it appears as though the K7S5A cannot supply a voltage as low as 1.45v - which isn't a problem if you are overclocking - but in my case I just want to run at stock speed and have a quiet system so that may be the limiting factor
I'm still researching though so there may be hope yet! Makes for some very interesting reading though
EDIT: My mistake - is does support 1.45v Its just that it doesn't allow modification once the default CPU voltage is detected - no issue for me as I dont plan to overclock...woohoo!
-
- Posts: 90
- Joined: Mon Feb 09, 2004 4:16 pm
- Location: Greece
Auto detection of mobile bartons isn't as simple as it sounds.
You see , being a mobile cpu , it actually has 2 multis (6 & 14) and 2 voltages (1.45 & 1.57) to choose from when running in a laptop.
Your motherboard could choose either of each.
My Abit detects 6 multi , 1.57V , others detect 14 multi , 1.45V etc
If your mobo doesn't allow voltage and multi adjustment , you could have a problem (albeit one solvable via pinmods).
You see , being a mobile cpu , it actually has 2 multis (6 & 14) and 2 voltages (1.45 & 1.57) to choose from when running in a laptop.
Your motherboard could choose either of each.
My Abit detects 6 multi , 1.57V , others detect 14 multi , 1.45V etc
If your mobo doesn't allow voltage and multi adjustment , you could have a problem (albeit one solvable via pinmods).
-
- Posts: 49
- Joined: Thu Nov 13, 2003 3:59 pm
- Location: Toronto, Canada home of the no left turn sign
Just for the record my motherboard (ASUS A7V8X - KT400 chipset) BIOS can't identify the Mobile XP2500+ and therefore defaults the FSB to 100 MHz. As there is no PCI lock on the chipset I am not able to increase this number much. Neither does the board allow voltages below 1.575 so I also stuck with that number. Hence I am thinking of upgrading to an Abit NF7-S or AN7 as they are now around CDN $110-135.
Simple question, friends. What multipliers this mighty Barton 2500+ mobile does support?
Mine old Barton 2500+ need 1.9V core to run at 3200+ rating, archieved by running it at 200x11 (or was that 10.5?) but it also did not allow me higher multiplier that 12.5 for some reason?
This mobile thing allow my what multipliers to use?
...and yup, i hoping for less heat and running this one at 3200+ with only 1.65V
Mine old Barton 2500+ need 1.9V core to run at 3200+ rating, archieved by running it at 200x11 (or was that 10.5?) but it also did not allow me higher multiplier that 12.5 for some reason?
This mobile thing allow my what multipliers to use?
...and yup, i hoping for less heat and running this one at 3200+ with only 1.65V
-
- Posts: 90
- Joined: Mon Feb 09, 2004 4:16 pm
- Location: Greece
@trodas
The cpu's multiplier is a 5bit number , however most motherboards can only change 4bits , leaving the msb (most significant bit) unchanged.
That means that if your cpu has a multi up to (and including) 12.5x , you can't go over 12.5x or if the cpu has a multi over 13x , you can't go below 13x.
So in fact the problem you described , is caused by your motherboard.
The only motherboard , I know of , that gives access to the full 5bit range of multis , is the abit NF7/NF7-S (not sure about AN7).
So in your case , it comes to how the cpu will be identified by the mobo.
If it is identified as 6x multi , then you will have the below 12.5x range , if it is identified as 14x , then you will have the over 13x range.
Keep in mind however that you can change the msb of the multi , by using a pinmod (inserting a wire in the socket) , I have done it to the second PC (using an Epox 8RDA+) and it works like a charm.
The cpu's multiplier is a 5bit number , however most motherboards can only change 4bits , leaving the msb (most significant bit) unchanged.
That means that if your cpu has a multi up to (and including) 12.5x , you can't go over 12.5x or if the cpu has a multi over 13x , you can't go below 13x.
So in fact the problem you described , is caused by your motherboard.
The only motherboard , I know of , that gives access to the full 5bit range of multis , is the abit NF7/NF7-S (not sure about AN7).
So in your case , it comes to how the cpu will be identified by the mobo.
If it is identified as 6x multi , then you will have the below 12.5x range , if it is identified as 14x , then you will have the over 13x range.
Keep in mind however that you can change the msb of the multi , by using a pinmod (inserting a wire in the socket) , I have done it to the second PC (using an Epox 8RDA+) and it works like a charm.
This sounds like what I ran into trying to run it in an ASUS A7SC that I just purchased. It says the multiplier is locked at 6, and runs 1.56 volts. Stock speed for my 2400+ is now 600Mhz. Even with maximum overclock via the BIOS control I can only get it up to 995 Mhz, and I can't do anything with the voltage. Needless to say, it looks like I have been burned by the "not exactly supported" motherboard problem.apocalypse80 wrote:Auto detection of mobile bartons isn't as simple as it sounds.
You see , being a mobile cpu , it actually has 2 multis (6 & 14) and 2 voltages (1.45 & 1.57) to choose from when running in a laptop.
Your motherboard could choose either of each.
My Abit detects 6 multi , 1.57V , others detect 14 multi , 1.45V etc
If your mobo doesn't allow voltage and multi adjustment , you could have a problem (albeit one solvable via pinmods).
So what's the use?
Just a quick thought.
I have a week 43 XP 2500+ at 200x11 at 1.6 vcore in bios, 1.58 in cpuz and 1.55 in MBM (couldn't tell you which is accurate). Anyway to me it seems that with the wide differences in results from the XP-M it is really a hit or miss as to how high you can get. The only positive that I see is the unlocked multiplier and maybe the undervolting/underclocking potential. This is basically only geared towards those people who already have a fairly recent/powerful cpu, if on the other hand you are looking to upgrade an older machine go for it and enjoy. I thought about going to the XP-M but with most people seemingly having to up the voltage to 1.7+ to get it to 2.3Ghz and up it doesn't seem to make sense when really you could do that with a regular 2500+ and the $25-40 that you would save would buy a better heatsink etc.
Andy
-----
XP 2500+ @ 200x11 @ 1.6 vcore
Abit NF7-S, Zalman 7000A modded with panaflo 92M @~9volts
ati 9800np/zalman passive cooler
I have a week 43 XP 2500+ at 200x11 at 1.6 vcore in bios, 1.58 in cpuz and 1.55 in MBM (couldn't tell you which is accurate). Anyway to me it seems that with the wide differences in results from the XP-M it is really a hit or miss as to how high you can get. The only positive that I see is the unlocked multiplier and maybe the undervolting/underclocking potential. This is basically only geared towards those people who already have a fairly recent/powerful cpu, if on the other hand you are looking to upgrade an older machine go for it and enjoy. I thought about going to the XP-M but with most people seemingly having to up the voltage to 1.7+ to get it to 2.3Ghz and up it doesn't seem to make sense when really you could do that with a regular 2500+ and the $25-40 that you would save would buy a better heatsink etc.
Andy
-----
XP 2500+ @ 200x11 @ 1.6 vcore
Abit NF7-S, Zalman 7000A modded with panaflo 92M @~9volts
ati 9800np/zalman passive cooler
-
- Posts: 39
- Joined: Sat Oct 04, 2003 4:44 pm
- Location: Venice
- Contact:
apocalypse80 - aaargh! This sounds so bad... So my Epox 8RDA+ can't go over multiplier 12.5? WTH... And i tought that Epox mobos are supposed to be great for overclocking
So, this driving me to important question - what is the multiplier of AMD AthlonXP mobile 2500+ with barton core?
Since one customer mentioned that: "The stepping i got was an IQYHA 0405 SPFW. Currently the chip is at 12.5 X 200 = 2.5 ghz" - so i do reasonably believe that either you are wrong and something is limited into my CPU, or he have the Abit NF7 mobo, or the cpu multiplier is bellow the 13x ...
Im confused, pls advice. Multiplier is not mentioned into the newegg description at all:
Manufacturer: AMD
Model: AXMH2500FQQ4C
Core: Barton
Operating Frequency: 1.83GHz
FSB: 266MHz
Cache: L1/64K+64K; L2/512K
Voltage: 1.45V
Process: 0.13Micron
Socket: Socket A
Multimedia Instruction: MMX, SSE, 3DNOW!, 3DNOW!+
Packaging: OEM
...ah!!! I got it! 266Mhz is 133MHz FSB and is i divide 1.83Ghz (1874) by 133 i got the multiplier - 14
And now what?
Can i lover the multiplier on Epox 8RDA+ mobo, or it require additional changes? Or all the users out there having Abit NF7? (somewhat i strongly dubt about...)
So, this driving me to important question - what is the multiplier of AMD AthlonXP mobile 2500+ with barton core?
Since one customer mentioned that: "The stepping i got was an IQYHA 0405 SPFW. Currently the chip is at 12.5 X 200 = 2.5 ghz" - so i do reasonably believe that either you are wrong and something is limited into my CPU, or he have the Abit NF7 mobo, or the cpu multiplier is bellow the 13x ...
Im confused, pls advice. Multiplier is not mentioned into the newegg description at all:
Manufacturer: AMD
Model: AXMH2500FQQ4C
Core: Barton
Operating Frequency: 1.83GHz
FSB: 266MHz
Cache: L1/64K+64K; L2/512K
Voltage: 1.45V
Process: 0.13Micron
Socket: Socket A
Multimedia Instruction: MMX, SSE, 3DNOW!, 3DNOW!+
Packaging: OEM
...ah!!! I got it! 266Mhz is 133MHz FSB and is i divide 1.83Ghz (1874) by 133 i got the multiplier - 14
And now what?
Can i lover the multiplier on Epox 8RDA+ mobo, or it require additional changes? Or all the users out there having Abit NF7? (somewhat i strongly dubt about...)
-
- Posts: 90
- Joined: Mon Feb 09, 2004 4:16 pm
- Location: Greece
I already wrote about that , the mobile 2500+ cpus actually have 2 default multis 6 & 14.
These two , if considered in binary form , differ only in the msb (the 5th bit that most mobos can't touch) , so in a laptop the cpu can work in 2 different frequencies 14x133 (at load) or 6x133 (at idle) , in order to remain cool , just by changing 1 bit .
Again it depends on autodetection , which of the 2 multis your mobo will detect.
The 8RDA+ in the second PC detects 6x multi , so I couldn't go over 12.5x multi , having bad ram (barely doing 166fsb) , that was unacceptable.
So I inserted a wire in the right pins to "set" the msb of the multi , clear CMOS and the cpu was detected as having a 14x multi (currently happy @ 15x160 = 2400).
The problem is , of course , that now I don't have access to multipliers below 13x , so if I got some good RAM for that PC , I would have to remove the pinmod to get the lower multis again.
Inserting a wire in the socket , really is no big deal , in the past all overclocking was done that way.
And here are the pinmods
I am succesfully (sp?) using the "high multis" on the second PC (8RDA+) and the "fsb 166" on the primary PC (NF7-S) , have never tried the other 2.
These two , if considered in binary form , differ only in the msb (the 5th bit that most mobos can't touch) , so in a laptop the cpu can work in 2 different frequencies 14x133 (at load) or 6x133 (at idle) , in order to remain cool , just by changing 1 bit .
Again it depends on autodetection , which of the 2 multis your mobo will detect.
The 8RDA+ in the second PC detects 6x multi , so I couldn't go over 12.5x multi , having bad ram (barely doing 166fsb) , that was unacceptable.
So I inserted a wire in the right pins to "set" the msb of the multi , clear CMOS and the cpu was detected as having a 14x multi (currently happy @ 15x160 = 2400).
The problem is , of course , that now I don't have access to multipliers below 13x , so if I got some good RAM for that PC , I would have to remove the pinmod to get the lower multis again.
Inserting a wire in the socket , really is no big deal , in the past all overclocking was done that way.
And here are the pinmods
I am succesfully (sp?) using the "high multis" on the second PC (8RDA+) and the "fsb 166" on the primary PC (NF7-S) , have never tried the other 2.
Great! Finally I got my very own Mobile XP 2600+. After some trial and error I got it running at 2GHz with FSB 166 and multiplier of 12. My motherboard (Asus A7N8X, I think ) does not allow much changing of parameters but at least I am free to choose the FSB without any wire mods from 100/133/166/200 and set the multiplier. Unfortunately, the RAM does not make it through the RAM test at 200 MHz (although it is rated as PC3200 ). Anyway, now I am running CPUburn at 50 degrees which is very low for me (30 degrees above ambient) and that is at 2 GHz! I think it would be nice to run at lower Vcore but I am a bit scared of making those irreversible mods and this mobo does not allow for undervolting. Anyway, this CPU is awesome I think I may reconsider all future plans and go always with mobile.
An M-Barton 2500, running at 2167mhz (the speed of a 3000) will put out 62.3W's, if its running at its stock voltage of 1.45. A desktop 2500, at stock speeds and voltage puts out 68.3W's. If you need to, you could overvolt the M-Barton up to 1.525v, and at 2167mhz it will put out 68.9W's, about the same as the desktop 2500.mpteach wrote:IM looking to run standard 3000 speeds with 2500 level wattage. what Mobile barton chip should i choose?
You'll be getting XP3000 performance, at XP2500 heat output, and it'll cost 30% less than buying a 3000! Not a bad deal, all around.
-
- Posts: 90
- Joined: Mon Feb 09, 2004 4:16 pm
- Location: Greece
Keep in mind that the 10x and 10.5x multipliers don't work very well on most nforce2 motherboards.
In fact , some motherboards don't even have them in the bios.
Also cas3 can mess with most nforce2 mobos (mine doesn't even boot with cas3).
@fabre , try a small multiplier (like 9x) to find the maximum fsb you motherboard/memory will handle , then work your way up.
Also try overvolting the memory a little bit , you may be able to use cas2.5.
In fact , some motherboards don't even have them in the bios.
Also cas3 can mess with most nforce2 mobos (mine doesn't even boot with cas3).
@fabre , try a small multiplier (like 9x) to find the maximum fsb you motherboard/memory will handle , then work your way up.
Also try overvolting the memory a little bit , you may be able to use cas2.5.
Does the Nf7-s support 10.5? i was planing on running 10.5 x 220. For what its worth i would use either 2 or 2.5 CAS.apocalypse80 wrote:Keep in mind that the 10x and 10.5x multipliers don't work very well on most nforce2 motherboards.
In fact , some motherboards don't even have them in the bios.
Also cas3 can mess with most nforce2 mobos (mine doesn't even boot with cas3).
@fabre , try a small multiplier (like 9x) to find the maximum fsb you motherboard/memory will handle , then work your way up.
Also try overvolting the memory a little bit , you may be able to use cas2.5.
-
- Posts: 90
- Joined: Mon Feb 09, 2004 4:16 pm
- Location: Greece
@fabre , there should be an option in the bios that will allow you to change the fsb in 1mhz increments.
Something like "system performance" or something , all nforce2 motherboards have something like that.
Yeah , the NF7-S supports 10x and 10.5x , but they din't work properly , until I flashed a tictac modded bios.
That is , by relaxing memory timings to 2.5-3-3-7 , I could get 239fsb with every multi , except 10x and 10.5x that wouldn't even boot over 222 fsb.
After the modded bios , they all work like a charm.
Also my board will not even boot with some multis, at any respectable fsb , unless I have cpu interface:enabled.
I always have it enabled anyway (boosts performance) , but it's definitely a glitch.
Something like "system performance" or something , all nforce2 motherboards have something like that.
Yeah , the NF7-S supports 10x and 10.5x , but they din't work properly , until I flashed a tictac modded bios.
That is , by relaxing memory timings to 2.5-3-3-7 , I could get 239fsb with every multi , except 10x and 10.5x that wouldn't even boot over 222 fsb.
After the modded bios , they all work like a charm.
Also my board will not even boot with some multis, at any respectable fsb , unless I have cpu interface:enabled.
I always have it enabled anyway (boosts performance) , but it's definitely a glitch.