Re: High end server and storage for a PostgreSQL OLTP system - Mailing list pgsql-performance

From William Yu
Subject Re: High end server and storage for a PostgreSQL OLTP system
Date
Msg-id 4201195B.6000203@talisys.com
Whole thread Raw
In response to Re: High end server and storage for a PostgreSQL OLTP system  (Bruce Momjian <pgman@candle.pha.pa.us>)
Responses Re: High end server and storage for a PostgreSQL OLTP system  (Bruce Momjian <pgman@candle.pha.pa.us>)
List pgsql-performance
>>You can get 64-bit Xeons also but it takes hit in the I/O department due
>>to the lack of a hardware I/O MMU which limits DMA transfers to
>>addresses below 4GB. This has a two-fold impact:
>>
>>1) transfering data to >4GB require first a transfer to <4GB and then a
>>copy to the final destination.
>>
>>2) You must allocate real memory 2X the address space of the devices to
>>act as bounce buffers. This is especially problematic for workstations
>>because if you put a 512MB Nvidia card in your computer for graphics
>>work -- you've just lost 1GB of memory. (I dunno how much the typical
>>SCSI/NIC/etc take up.)
>
>
> I thought Intel was copying AMD's 64-bit API.  Is Intel's
> implementation as poor as you description?  Does Intel have any better
> 64-bit offering other than the Itanium/Itanic?

Unfortunately, there's no easy way for Intel to have implemented a
64-bit IOMMU under their current restrictions. The memory controller
resides on the chipset and to upgrade the functionality significantly,
it would probably require changing the bus protocol. It's not that they
couldn't do it -- it would just require all Intel chipset/MB
vendors/partners to go through the process of creating & validating
totally new products. A way lengthier process than just producing 64-bit
CPUs that drop into current motherboards.

pgsql-performance by date:

Previous
From: Tom Lane
Date:
Subject: Re: Bad query optimizer misestimation because of TOAST tables
Next
From: Markus Schaber
Date:
Subject: Re: Bad query optimizer misestimation because of TOAST