A8V - Rev 2 vs Rev 1

[CapeGuy]

When I bought my A8V from NewEgg in mid-January, I received
a Rev 1 board. Subsequently, I heard that Rev 2 boards had been
shipping since November or so. Am I missing anything significant
by having the older Rev 1 board? Also, is it typical for NewEgg to
ship "out of date" stuff without letting the customer know?

 

[Paul]

When I bought my A8V from NewEgg in mid-January, I received
a Rev 1 board. Subsequently, I heard that Rev 2 boards had been
shipping since November or so. Am I missing anything significant
by having the older Rev 1 board? Also, is it typical for NewEgg to
ship "out of date" stuff without letting the customer know?

AGP/PCI lock is officially supported on Rev.2

Paul

 

[RonK]

Hey Paul, What is AGP/PCI lock and what does it do ?
Thanks
Ron

 

[sean]

I would send it back and get the latest release of the board. It is fair
to expect that when you buy a new motherboard, it should be the latest
release, especially when it has been out for a few months. NewEgg are
trying to flogg their old stock to you.

Sean

 

[Paul]

Hey Paul, What is AGP/PCI lock and what does it do ?
Thanks
Ron

An AGP/PCI lock keeps the clocks at 66.66/33.33 MHz, while
you are overclocking. On previous generations of boards lacking
this feature, as you raised the FSB frequency, the AGP and
PCI would go up also. Once PCI, which is normally 33MHz, goes
above about 37.5MHz, you could corrupt an IDE disk drive, as
they did the timing on the drive interface with the PCI clock.
(Not all boards did that, just the stupid ones And if it
wasn't the IDE interface, then at perhaps 40MHz, the rest of
the PCI bus would break.)

To make a working AGP/PCI lock requires a couple of things.
It needs the chipset to be designed with asynchronous clocking.
That means normally one clock signal would enter a certain
part of the chip, and for a lock, you need a separate clock for
two parts of the circuit. Special precautions must be made
to pass data between the two domains. (I suspect these requirements
weren't met properly on the chipset of the A8V rev one, and
Asus had to disable the lock on the rev one, to avoid problems
with the chipset. It was fixed on rev two. If you are not an
overclocker, the lock is a "don't care" feature you will never
use.)

The second requirement, is a flexible clockgen chip, that can
hold the AGP/PCI clock at a constant frequency, while changing
the FSB.

Even with all this fancy mumbo-jumbo, the new boards do have
problems with the SATA interface. It seems mobo designers
learn the same lessons over and over again, as there are some
SATA interfaces on motherboards that malfunction when you
overclock past a certain point. That is why it pays to read
the reviews for your motherboard on Anandtech, as they sometimes
mention which interfaces can suffer from corruption.
Seasoned overclockers pick their disk drive interface with
great care, due to the potential to lose the contents of the
disk drive while experimenting.

HTH,
Paul

 

[RonK]

Thanks a lot for the info Paul.
My A8V is revision 1 but I have had no problems running it at 10% overclock.

 

[Highlandish]

An AGP/PCI lock keeps the clocks at 66.66/33.33 MHz, while
you are overclocking. On previous generations of boards lacking
this feature, as you raised the FSB frequency, the AGP and
PCI would go up also. Once PCI, which is normally 33MHz, goes
above about 37.5MHz, you could corrupt an IDE disk drive, as
they did the timing on the drive interface with the PCI clock.
(Not all boards did that, just the stupid ones And if it
wasn't the IDE interface, then at perhaps 40MHz, the rest of
the PCI bus would break.)

To make a working AGP/PCI lock requires a couple of things.
It needs the chipset to be designed with asynchronous clocking.
That means normally one clock signal would enter a certain
part of the chip, and for a lock, you need a separate clock for
two parts of the circuit. Special precautions must be made
to pass data between the two domains. (I suspect these requirements
weren't met properly on the chipset of the A8V rev one, and
Asus had to disable the lock on the rev one, to avoid problems
with the chipset. It was fixed on rev two. If you are not an
overclocker, the lock is a "don't care" feature you will never
use.)

The second requirement, is a flexible clockgen chip, that can
hold the AGP/PCI clock at a constant frequency, while changing
the FSB.

Even with all this fancy mumbo-jumbo, the new boards do have
problems with the SATA interface. It seems mobo designers
learn the same lessons over and over again, as there are some
SATA interfaces on motherboards that malfunction when you
overclock past a certain point. That is why it pays to read
the reviews for your motherboard on Anandtech, as they sometimes
mention which interfaces can suffer from corruption.
Seasoned overclockers pick their disk drive interface with
great care, due to the potential to lose the contents of the
disk drive while experimenting.

HTH,
Paul

I'm about to manually OC my rev 1 with a 939 3000+ to a 3800+. is this going
to effect my ide/sata bus?

 

[Paul]

I'm about to manually OC my rev 1 with a 939 3000+ to a 3800+. is this going
to effect my ide/sata bus?

That is going from 1800MHz to 2400MHz. A 33% overclock.

I'd take a look through this thread. Early in the thread,
some people are using a modded BIOS, and it shows an item
for the lock. They seem to be able to run high clocks on
the FSB (but they are also careful as to how they connect
a hard disk). This thread is too long for me to read the
whole thing.

http://www.abxzone.com/forums/showthread.php?t=74370

I think you really need to find an overclocking summary
thread. This link was posted in this group a few days
ago:

http://www.ocforums.com/showthread.php?t=332591

The K8V is a board that doesn't have a working lock, and
it hits a "wall" at 237MHz. If the A8V Rev1 is going to go
farther than that, maybe the modded BIOS will help. So far
I haven't read a thread, where someone tries to overclock
a Rev.1 in its "out of the box" state. Sounds like plenty
of experiments to come...

HTH,
Paul

 

[CapeGuy]

If I'm not interested in overclocking, is there any
other reason that Rev 2 may be better than Rev 1?

 

[Highlandish]

That is going from 1800MHz to 2400MHz. A 33% overclock.

I'd take a look through this thread. Early in the thread,
some people are using a modded BIOS, and it shows an item
for the lock. They seem to be able to run high clocks on
the FSB (but they are also careful as to how they connect
a hard disk). This thread is too long for me to read the
whole thing.

http://www.abxzone.com/forums/showthread.php?t=74370

I think you really need to find an overclocking summary
thread. This link was posted in this group a few days
ago:

http://www.ocforums.com/showthread.php?t=332591

The K8V is a board that doesn't have a working lock, and
it hits a "wall" at 237MHz. If the A8V Rev1 is going to go
farther than that, maybe the modded BIOS will help. So far
I haven't read a thread, where someone tries to overclock
a Rev.1 in its "out of the box" state. Sounds like plenty
of experiments to come...

HTH,
Paul

thanks, I read the latter mentioned link to oc the a8v, but the first link
was great, lots of info there too. little question, I currently have the
1009 bios, is that pci locked? will it hurt to go back to the modded 1005.21
bios?

 

[Wayne Fulton]

When I bought my A8V from NewEgg in mid-January, I received
a Rev 1 board. Subsequently, I heard that Rev 2 boards had been
shipping since November or so. Am I missing anything significant
by having the older Rev 1 board? Also, is it typical for NewEgg to
ship "out of date" stuff without letting the customer know?

I bought an A8V from NewEgg in early December, with 1008 BIOS.
It says Rev 2.0 printed on the motherboard.

However CPUZ and Asus Probe say Rev 1.xx.
Sounds like something in BIOS is not updated ?

 

[Paul]

thanks, I read the latter mentioned link to oc the a8v, but the first link
was great, lots of info there too. little question, I currently have the
1009 bios, is that pci locked? will it hurt to go back to the modded 1005.21
bios?

If you look at the cpusupport web page, 1005 isn't even on the
radar, so I don't know what will happen if you flash back to
that old a BIOS. If you had an even older BIOS than that, to
start with, and were able to boot, it is probably safe to use
1005. But there are likely stability improvements in the later
BIOS, so it is a tough call.

In terms of what entry to expect in the BIOS screen, I wasn't
able to find a picture of an A8V screen with the lock setting.
If you look at the A8V-E PDF user manual, in Advanced:Frequency
Configuration, it has:

PCI clock sync to CPU [Enabled]
PCI clock 33.0MHz

The [Enabled] setting means "not locked". Basically what
happens, is the unchangable value below the control, is
a simple divide from the FSB, like 198/6=33.0 . As you increase
the clock from 200 to 227, the PCI clock will rise to 37.8MHz.

If you set that setting to [Disabled], then you are locked. The
PCI clock setting below should become [33.0MHz] , where the
bracket is indicating that you can change the value. Generally
only a few options would be available, like [Auto,33.0,37.5MHz].
A setting of 33.0 is the best choice (not auto, because sometimes
BIOS do stupid things on auto).

I'm hoping the A8V-E BIOS is based on the A8V BIOS, when I suggest
that is what the interface looks like. Asus never updates the manual
if the BIOS feature set changes, so I cannot expect to ever see
a manual page for how this would look on the A8V.

If, in the 1009 BIOS, you have the above two entries or something
similar, then you likely don't need 1005.

When I was browsing Abxzone, I saw a couple of comments (URLs
got lost along the way). One was, disk corruption off the Via
interface at about 228MHz. And, if somehow you can get the lock
to work, there was a comment that you can lift HT to 280MHz on
a rev 1.

Even if the BIOS shows the above entry, there is really no way
to know whether it is working or not. A PCI card called the
"PC Geiger", a product from ioss.com.tw, can actually measure the
PCI clock rate, but those cards are not widely available.

You might get a copy of this tool. There is a version for
the A8V, and even if you choose not to use it to set the clock
rates on the board, it is possible the interface will identify
the current value of the PCI clock. If it reads 33.0 MHz no
matter what HT(FSB) clock you select, then the lock must be
working.

http://www.cpuid.com/clockgen.php

HTH,
Paul

 

[Highlandish]

thanks, I read the latter mentioned link to oc the a8v, but the
first link was great, lots of info there too. little question, I
currently have the 1009 bios, is that pci locked? will it hurt to go
back to the modded 1005.21 bios?

If you look at the cpusupport web page, 1005 isn't even on the
radar, so I don't know what will happen if you flash back to
that old a BIOS. If you had an even older BIOS than that, to
start with, and were able to boot, it is probably safe to use
1005. But there are likely stability improvements in the later
BIOS, so it is a tough call.

In terms of what entry to expect in the BIOS screen, I wasn't
able to find a picture of an A8V screen with the lock setting.
If you look at the A8V-E PDF user manual, in Advanced:Frequency
Configuration, it has:

PCI clock sync to CPU [Enabled]
PCI clock 33.0MHz

The [Enabled] setting means "not locked". Basically what
happens, is the unchangable value below the control, is
a simple divide from the FSB, like 198/6=33.0 . As you increase
the clock from 200 to 227, the PCI clock will rise to 37.8MHz.

If you set that setting to [Disabled], then you are locked. The
PCI clock setting below should become [33.0MHz] , where the
bracket is indicating that you can change the value. Generally
only a few options would be available, like [Auto,33.0,37.5MHz].
A setting of 33.0 is the best choice (not auto, because sometimes
BIOS do stupid things on auto).

I'm hoping the A8V-E BIOS is based on the A8V BIOS, when I suggest
that is what the interface looks like. Asus never updates the manual
if the BIOS feature set changes, so I cannot expect to ever see
a manual page for how this would look on the A8V.

If, in the 1009 BIOS, you have the above two entries or something
similar, then you likely don't need 1005.

When I was browsing Abxzone, I saw a couple of comments (URLs
got lost along the way). One was, disk corruption off the Via
interface at about 228MHz. And, if somehow you can get the lock
to work, there was a comment that you can lift HT to 280MHz on
a rev 1.

Even if the BIOS shows the above entry, there is really no way
to know whether it is working or not. A PCI card called the
"PC Geiger", a product from ioss.com.tw, can actually measure the
PCI clock rate, but those cards are not widely available.

You might get a copy of this tool. There is a version for
the A8V, and even if you choose not to use it to set the clock
rates on the board, it is possible the interface will identify
the current value of the PCI clock. If it reads 33.0 MHz no
matter what HT(FSB) clock you select, then the lock must be
working.

http://www.cpuid.com/clockgen.php

HTH,
Paul

thanks again