increasing FSB with Thunderbird

What kind of performance increase would I get if I took my Athlon 850
from 100x8.5 / 850MHz to 146x6 / 876MHz? I've got Legend CAS3-3-3 SDRAM.

On Fri, 30 Jan 2004 01:40:05 +1300, Species wrote:

> What kind of performance increase would I get if I took my Athlon 850
> from 100x8.5 / 850MHz to 146x6 / 876MHz? I've got Legend CAS3-3-3 SDRAM.

Not enough to notice even if it worked. With a 146Mhz FSB you run into
concerns for ram speed, PCI and AGP bus speeds. How fast is your ram?
PC133 is rated only for 133Mhz. I'd just up the multiplier to get more
speed, or maybe change the FSB to 133MHz as long as I knew the ram would
work there. 7.5x133 should be doable, or 10x100. May get higher, just
depends on your components. I had my old 1gig tbird up over 1400MHz.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.html

Wes Newell wrote:

> On Fri, 30 Jan 2004 01:40:05 +1300, Species wrote:
>
> > What kind of performance increase would I get if I took my Athlon 850
> > from 100x8.5 / 850MHz to 146x6 / 876MHz? I've got Legend CAS3-3-3 SDRAM.
>
> Not enough to notice even if it worked.

Hmm okay. Maybe I'll get a ThoroughbredB when they're being given away

> With a 146Mhz FSB you run into
> concerns for ram speed, PCI and AGP bus speeds.

They all cope fine with the PCI at 37MHz. I wonder why components today have
such trouble with high PCI speeds, when back in the VESA days lots of things
ran fine at 40 or even 50MHz.

> How fast is your ram?
> PC133 is rated only for 133Mhz.

That's true, but this module is a more recent part and copes pretty well with
the higher speeds.

> I'd just up the multiplier to get more
> speed,

Stupid motherboard doesn't have any multiplier settings I'd have to get a
converter.

> or maybe change the FSB to 133MHz as long as I knew the ram would
> work there. 7.5x133 should be doable, or 10x100. May get higher, just
> depends on your components. I had my old 1gig tbird up over 1400MHz.

I've got no doubt the CPU could beat 1GHz, but the motherboard doesn't have
any vcore adjustment, and it is limited to how far it can go with the
defaults.

Asus has got some pretty weird definitions of "easy jumper-free overclocking".
No vcore nor no multiplier adjustments. How're you meant to overclock easily
when you can't move anything other than the FSB? I think Asus is
overrated.

> --
> Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
> http://mysite.verizon.net/res0exft/cpu.html

That's pretty good. Maybe you could run it at 22x110FSB?

It is good to see you didn't simply chuck out your motherboard like most
people would when they get a new CPU.

On Fri, 30 Jan 2004 10:09:08 +1300, Species wrote:

> I've got no doubt the CPU could beat 1GHz, but the motherboard doesn't have
> any vcore adjustment, and it is limited to how far it can go with the
> defaults.
>
> Asus has got some pretty weird definitions of "easy jumper-free overclocking".
> No vcore nor no multiplier adjustments. How're you meant to overclock easily
> when you can't move anything other than the FSB? I think Asus is
> overrated.
>
You can change vcore and or multiplier using the pinmod. Just a little
wire.

http://www.ocinside.de/go_e.html?/ht...md_pinmod.html

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.html

Wes Newell wrote:

> On Fri, 30 Jan 2004 10:09:08 +1300, Species wrote:
>
> > I've got no doubt the CPU could beat 1GHz, but the motherboard doesn't have
> > any vcore adjustment, and it is limited to how far it can go with the
> > defaults.
> >
> > Asus has got some pretty weird definitions of "easy jumper-free overclocking".
> > No vcore nor no multiplier adjustments. How're you meant to overclock easily
> > when you can't move anything other than the FSB? I think Asus is
> > overrated.
> >
> You can change vcore and or multiplier using the pinmod. Just a little
> wire.

Thanks, I didn't know that.

> http://www.ocinside.de/go_e.html?/ht...md_pinmod.html
>

That seems to be about Athlon XPs and not Thunderbirds. Does that information apply
to Thunderbirds too? It doesn't say.

>
> --
> Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
> http://mysite.verizon.net/res0exft/cpu.html

On Sat, 31 Jan 2004 03:12:56 +1300, Species wrote:

> Wes Newell wrote:
>
>> You can change vcore and or multiplier using the pinmod. Just a little
>> wire.
>
> Thanks, I didn't know that.
> That seems to be about Athlon XPs and not Thunderbirds. Does that
> information apply to Thunderbirds too? It doesn't say.
>
The vcore pinmod would be the same. And all but the 5th multiplier bit
(pin AJ27) would be the same. You probably could connect it as it's not
used on the Tbird. For vcore, you actually only need to add 1 wire to set
the .1v bit to get it up to 1.85v. And normally you won't ned all 4 wires
to set the multiplier you want either, although it may be easier than
figuring out which ones you do need. This can also be accomplished on the
cpu itself if you want to take that route. Or you could wire up jumper
blocks/switches to the back of the MB if you want easier ways to change
all this. Options are many. As an example, let's say you want to test with
1.85v. you could wire up a spst or spdt switch to the .1v control pin and
then turn it off and on without having to take out the cpu. Many many ways
to skin a cat.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.html

Wes Newell wrote:

> On Sat, 31 Jan 2004 03:12:56 +1300, Species wrote:
>
> > Wes Newell wrote:
> >
> >> You can change vcore and or multiplier using the pinmod. Just a little
> >> wire.
> >
> > Thanks, I didn't know that.
> > That seems to be about Athlon XPs and not Thunderbirds. Does that
> > information apply to Thunderbirds too? It doesn't say.
> >
> The vcore pinmod would be the same. And all but the 5th multiplier bit
> (pin AJ27) would be the same. You probably could connect it as it's not
> used on the Tbird. For vcore, you actually only need to add 1 wire to set
> the .1v bit to get it up to 1.85v. And normally you won't ned all 4 wires
> to set the multiplier you want either, although it may be easier than
> figuring out which ones you do need. This can also be accomplished on the
> cpu itself if you want to take that route. Or you could wire up jumper
> blocks/switches to the back of the MB if you want easier ways to change
> all this. Options are many. As an example, let's say you want to test with
> 1.85v. you could wire up a spst or spdt switch to the .1v control pin and
> then turn it off and on without having to take out the cpu. Many many ways
> to skin a cat.

does the CPU have to be unlocked before the pin mod will work?

> --
> Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
> http://mysite.verizon.net/res0exft/cpu.html

On Tue, 03 Feb 2004 20:20:20 +1300, Species wrote:

> does the CPU have to be unlocked before the pin mod will work?
>
For the multiplier, yes. Not for vcore.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.html

Wes Newell wrote:

> On Tue, 03 Feb 2004 20:20:20 +1300, Species wrote:
>
> > does the CPU have to be unlocked before the pin mod will work?
> >
> For the multiplier, yes. Not for vcore.

That'd be done by cutting the L1 bridges wouldn't it? I think I'll set a
6.5x or 7x multiplier.

I read that POSTing can be problematic when running it with the 133MHz+
bus, as it uses the original (8.5x) multiplier momentarily, is this
true?

On Wed, 04 Feb 2004 19:45:22 +1300, Species wrote:

> Wes Newell wrote:
>
>> On Tue, 03 Feb 2004 20:20:20 +1300, Species wrote:
>>
>> > does the CPU have to be unlocked before the pin mod will work?
>> >
>> For the multiplier, yes. Not for vcore.
>
> That'd be done by cutting the L1 bridges wouldn't it? I think I'll set a
> 6.5x or 7x multiplier.
>
No. The L1 bridges are not cut on unlocked cpu's. All Palomino core cpu's
had them cut and you had to reconnect them to unlock the cpu. L1'a are not
cut on Tbreds/Bartons, but newer cpu's are locked another way internally
and the condiditon of the L1's doesn't matter.

> I read that POSTing can be problematic when running it with the 133MHz+
> bus, as it uses the original (8.5x) multiplier momentarily, is this true?

Depends on Mb. Only a very small amount of boards have this problem afaik.
And pressing reset does boot the system.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.html