Is it possible to upgrade the Pentium 90 chip on an F330? (Haven't opened mine in a while... Is it in a ZIF socket?)
Has anyone tried or looked into this?
slot 0: System Board 90 MHz (NetApp System Board I rev-b) Memory Size: 256 MB
I know it is possible to upgrade the entire head or move the disks to another array with the same size disks.
Dan
Is it possible to upgrade the Pentium 90 chip on an F330? (Haven't opened mine in a while... Is it in a ZIF socket?)
Has anyone tried or looked into this?
Yes you can, and we (network appliance) have done it up to 155 unfortunatly the bottlneck is the memory subsystem which can't be upgraded so you will get the same performance while running a non-standard config (we ran 155 to see if it helped with performance we have not run it through a qa cycle). So there really is NO reason for this. Remember all the cpu is doing is moving bytes with the ocasional XOR or checksum (still memory bound).
Sean O'Malley Network Appliance Corp 2770 San Tomas Expressway Santa Clara, CA 95051 Phone: 408 367-3266 Fax: 408 367 3151 Email: sean@netapp.com
slot 0: System Board 90 MHz (NetApp System Board I rev-b) Memory Size: 256 MBI know it is possible to upgrade the entire head or move the disks to another array with the same size disks.
Dan
-- Daniel Quinlan (at work) Linux, our last best hope for Unix quinlan@transmeta.com http://www.pathname.com/~quinlan/ ... (Got BOFH? Looking for good ones.)
On Tue, 9 Sep 1997, Sean W. O'Malley wrote:
|> |> Is it possible to upgrade the Pentium 90 chip on an F330? (Haven't |> opened mine in a while... Is it in a ZIF socket?) |> |> Has anyone tried or looked into this? | |Yes you can, and we (network appliance) have done it up to 155 unfortunatly |the bottlneck is the memory subsystem which can't be upgraded so |you will get the same performance while running a non-standard |config (we ran 155 to see if it helped with performance we have |not run it through a qa cycle). So there really is NO reason for |this. Remember all the cpu is doing is moving bytes with the |ocasional XOR or checksum (still memory bound).
How would one go about upgrading the cpu in a F220 then?
Jonah
Jonah Barron Yokubaitis | Austin|San Antonio|Houston President | Dallas|Fort Worth|Boerne Texas.Net | Georgetown|Dripping Springs http://www.texas.net | Making 56k affordable
I seemed to have unintentionally opened a can of worms here in trying to stop people from trying to upgrade to higher clock speed pentiums. The offcially netapp policy is:
1) we don't recommend it as it doesn't provide any performance gain 2) we don't support such modifications - you may be voiding their warranty 3) we discourage this sort of tinkering - the previous note was just mentioning what I saw in the lab, and wasn't encouraging this sort of activity. 4) No such beast has ever gone through QA and you DON'T want to run a configuation with a non-standard nor QA'ed mother board.
The offical Sean O'Malley policy is buy a 520...I spent a good three or four months trying to make the pentium boxes go faster and if I had suceeded there would have been an F340. Messing around with the pentium boxes is a waste of time. If you have a real performance problem with any of the pentium based machine the 520 is the answer. Or get another box and split the load.
Sean O'Malley
On Tue, 9 Sep 1997, Sean W. O'Malley wrote:
|> |> Is it possible to upgrade the Pentium 90 chip on an F330? (Haven't |> opened mine in a while... Is it in a ZIF socket?) |> |> Has anyone tried or looked into this? | |Yes you can, and we (network appliance) have done it up to 155 unfortunatly |the bottlneck is the memory subsystem which can't be upgraded so |you will get the same performance while running a non-standard |config (we ran 155 to see if it helped with performance we have |not run it through a qa cycle). So there really is NO reason for |this. Remember all the cpu is doing is moving bytes with the |ocasional XOR or checksum (still memory bound).
How would one go about upgrading the cpu in a F220 then?
Jonah
Jonah Barron Yokubaitis | Austin|San Antonio|Houston President | Dallas|Fort Worth|Boerne Texas.Net | Georgetown|Dripping Springs http://www.texas.net | Making 56k affordable
On Wed, 10 Sep 1997, Sean W. O'Malley wrote:
| |I seemed to have unintentionally opened a can of worms here in |trying to stop people from trying to upgrade to higher clock |speed pentiums. The offcially netapp policy is: | | 1) we don't recommend it as it doesn't provide any performance gain | 2) we don't support such modifications - you may be voiding their | warranty | 3) we discourage this sort of tinkering - the previous note was just | mentioning what I saw in the lab, and wasn't encouraging this | sort of activity. | 4) No such beast has ever gone through QA and you DON'T want to run | a configuation with a non-standard nor QA'ed mother board. | |The offical Sean O'Malley policy is buy a 520...I spent a good three or |four months trying to make the pentium boxes go faster and if I had |suceeded there would have been an F340. Messing around with the pentium |boxes is a waste of time. If you have a real performance problem |with any of the pentium based machine the 520 is the answer. Or get |another box and split the load. |
good advice Sean. Did you testing engineers ever try Pentium PROs? We have found on our NT servers that PPROs have a _huge_ IO advantage over Pentiums because of the L2 cache running at cpu clock. The memory bus of a PPRO is still 64bits though. What is the memory bus on the Alpha based Netapps? We just submitted a PO for a F630 and are curious as to what its memory bus is.
Jonah
Jonah Barron Yokubaitis | Austin|San Antonio|Houston President | Dallas|Fort Worth|Boerne Texas.Net | Georgetown|Dripping Springs http://www.texas.net | Making 56k affordable
-
Daniel Quinlan -
Jonah Yokubaitis -
sean＠netapp.com