Can there be in whatever way to accelerate bulk placing in 2005? I am considering partitioning, could it be good?

See Loading Bulk Data into a Partitioned Table for any discussion of Bulk Place and partitioning interaction. While partitioning doesn't accelerate bulk place, it enables you to definitely bulk place parallel loads into separate unpartitioned tables after which use ALTER TABLE ... SWITH PARTITION to change all submitted right into a single partitioned table, see Transferring Data Efficiently by Using Partition Switching. However to get this done it's needed the information is correctly partitioned within the ETL process first, to ensure that is distributed towards the proper bulk place load.

Generally, likely to such extremes is not required. You need to have the ability to obtain impressive bulk place speeds in simple situations w/o problems, you sure you squashed everything out of the plain ole' bulk place? For example, are you certain that bulk-logging happens? And never last, is the disk fast (both LDF and MDF)? Shall we be speaking about 20 spindles in removed RAID 10, or perhaps is a 5400 Revoltions per minute 1TB slow-as-it-will get disk?

Here is a similar question which addresses exactly the same problem to some extent:

http://stackoverflow.com/questions/24200/whats-the-fastest-way-to-bulk-insert-a-lot-of-data-in-sql-server-c-client