engine and books on usb stick?

[ChessAddict]

Would it be okay to put my engine and books on a usb stick?

[Peter Skinner]

Not at all. I run all my chess software off my USB drive (500GB), or my 16 GB flash drive.

Just to see if it was slower and what not, I ran the Crafty benchmark and the results were identical.

Just make sure it is a fast flash drive. Some have horrible read/write speeds and that I am sure will make a difference.

Peter

[ZeroOne]

The engine is loaded into the main memory of the computer anyway so there's no difference. The book shouldn't make much difference either as accessing it is a pretty straight forward operation, I think. But putting endgame tablebases on a USB stick would improve their performance, as they are accessed pretty randomly and flash drives are a lot faster in that than regular disks. The downside is that you cannot fit many tablebases into a USB stick, but, say, you could fit all 5-men Nalimov bases into an 8 GB stick (which would cost something like $40...$60).

[BBauer]

Usb sticks may have a faster access (factor 10) time but throughput is usually slower than SATA. Best is putting table bases into memory. This can be done easyly in linux. So best is to have as much memory as possible.
regards
Bernhard

[bob]

Not at all. I run all my chess software off my USB drive (500GB), or my 16 GB flash drive.

Just to see if it was slower and what not, I ran the Crafty benchmark and the results were identical.

Just make sure it is a fast flash drive. Some have horrible read/write speeds and that I am sure will make a difference.

Peter

this is causing confusion. When you run the benchmark, you are not touching the USB stick or anything else. The stick is treated as a disk drive. The program is first loaded into memory from the stick, then it runs normally. If you had endgame tables on the stick, then it would influence execution speed. But not as you describe it. The only difference would be how long it takes to load from the stick vs loading from disks. For good disks, they will smoke the flash memory drives in that regard...

[bob]

Usb sticks may have a faster access (factor 10) time but throughput is usually slower than SATA. Best is putting table bases into memory. This can be done easyly in linux. So best is to have as much memory as possible.
regards
Bernhard

And SATA is far worse than good SCSI drives, particularly when using raid-0 to parallellize the reads/writes...

[BBauer]

You are right.
To my knowlege SATA gives up to 90 MB/sec.
But even if you get a factor of 4 for SCSI and raid0 this is still *slow* compared with memory access.
By the way, what happens if several threads try to access the same tb?
Does this leads to an even more significant slow down?
Here is a test position:

Study by Awerbach
FEN: 2k2K2/8/pp6/2p5/2P5/PP6/8/8 w - - 0 1

kind regards
Bernhard

[BBauer]

You are right.
To my knowlege SATA gives up to 90 MB/sec.
But even if you get a factor of 4 for SCSI and raid0 this is still *slow* compared with memory access.
By the way, what happens if several threads try to access the same tb?
Does this leads to an even more significant slow down?
Here is a test position:

Study by Awerbach
FEN: 2k2K2/8/pp6/2p5/2P5/PP6/8/8 w - - 0 1

kind regards
Bernhard

[Terry McCracken]

You are right.
To my knowlege SATA gives up to 90 MB/sec.
But even if you get a factor of 4 for SCSI and raid0 this is still *slow* compared with memory access.
By the way, what happens if several threads try to access the same tb?
Does this leads to an even more significant slow down?
Here is a test position:

Study by Awerbach
[d]2k2K2/8/pp6/2p5/2P5/PP6/8/8 w - - 0 1

kind regards
Bernhard

[Terry McCracken]

You are right.
To my knowlege SATA gives up to 90 MB/sec.
But even if you get a factor of 4 for SCSI and raid0 this is still *slow* compared with memory access.
By the way, what happens if several threads try to access the same tb?
Does this leads to an even more significant slow down?
Here is a test position:

Study by Awerbach
[d]2k2K2/8/pp6/2p5/2P5/PP6/8/8 w - - 0 1

kind regards
Bernhard

Nice position! It's deceptive. At first glance it would appear Ke8 wins, but it doesn't!

1. a4!! wins.

Terry