Question

Nuance pcache.*

Quick Fields

Updated April 23, 2015

asked on April 22, 2015 • Show version history

We have been using Quick Fields lately to process lots of discovery documents for a law firm. One of the big tasks is processing and OCR'ing a lot of documents. We have been looking into the performance of this process overall and it seems like we might be disk bound even though we are running with SSD's. The CPU never gets past 25-30%.

There is a file that seems to be written to very heavily called pcache.934 (at least during the instance that is running now). It is located in the c:\users\administrator\AppData\Local\Temp\1\Nuance folder.

Two questions:

1. Is QuickFields multi-threaded and if so why isn't it hitting 100% during the OCR process? [EDIT: Upon some additional research I discovered OCR is a single thread process]. Are there any solutions to this? Maybe the distributed computing option recently introduced?

2. How can I have the pcache file be on a different volume that would be a faster disk?

0 0

Replies

replied on April 23, 2015

You're right, the OCR process can only use one CPU. I'm guessing your machine has 4, which is why you're only seeing it hovering at 25% utilization. Distributed Processing will take advantage of the multiple CPUs by launching separate OCR tasks for different documents. Quick Fields image processing is serial since the order of pages is important when you're identifying documents, so by design, it does not do parallel processing for pages or documents.

For the second question, it's probably best you ask Nuance if there are any options for it, Laserfiche does not have any control over it. You could just move the Windows user's default temp folder to a different drive.

1 0

replied on April 23, 2015

I installed the DCC software today and generally it seems to work ok, but definitely a preview type of thing. There are several jobs that had failed, but so far unable to determine why and/or how to resubmit to attempt the processing again.

It is pretty useful how it enables OCR processing via the web access client. That provides a nice, easy way to have the OCR processing done via DCC.

I'm sure it will be much better in the release version. If it could create pages from PDF's and then OCR that would be ideal.

0 0

You are not allowed to follow up in this post.

Question

Question

Nuance pcache.*

Replies

Sign in to reply to this post.