Hardware resource and number of studies

I am running into extremely slow query response for dgate 1.4.17alpha with 2M images (46 min), especially when doing image count on all studies. My machine is Dell T5300, 64-bit, Dual core 2.6 GHz with only 4GB RAM, and I just about to upgrade to 20GB RAM. I have external 16TB software RAID-6 storage via eSATA for data and its performance is pretty good. OS is Fedora17. Performance started to slow down with increase in number of images, due to hardware resources.

Image count (Number of study related instances) query used (with dcmtk's findscu):
time findscu -aet MYAET -aec DGATE -S -k 0008,0052=STUDY -k 0020,000D -k 0020,1208 localhost 1234

With same OS, same dgate version, 2M images, and same external RAID storage, but with 8 core CPU and 16GB RAM, the same query response is about 6 seconds.

I would like to know what hardware resources are adequate for conquest with 2M images and up. Is there a hardware/DB/OS table for conquest put by users to take a look at, from performance standpoint?

Sundar

MySQL 5.5.32-1.fc17.x86_64. Meanwhile RAM was upgraded to 20GB, and still I run into slow query the first time.
/etc/my.cnf has these:
[mysqld]
innodb_buffer_pool_size=14G
query_cache_type = 1
query_cache_size = 256M

First time run of this query took 46min. While running top command showed:
CPU usage at 99% and RAM usage at 10%

Second time run took only 2 seconds, due to settings above. Memory upgrade did help for repeated queries.

If there is a dgate setting to improve the performance of query involving NumberOfStudyRelatedInstances, please let me know since this more of a CPU intensive task. Thanks.

Sundar

Sorry for the late update. This just turned out to be database issue, and regenerating it resolved it. Now it takes about 2.5sec which took 46min before, on the same machine.