BobWalraven
Technical User
Those of you who visit the peachtreeusers forums have seen my benchmark tests of Peachtree 2007 and Quantum performance. I have done a few more tests to measure the speed of 2007 and Quantum at reading records. The reason we expect a difference is because Peachtree has changed from Btrieve 6.15 to Pervasive 9.1 Btrieve for 2007 and for Quantum they have turned off Btrieve data compression, which should lead to somewhat faster reads.
The reason I looked at read timing and not write timing is because the vast majority of operations that are performed by Peachtree are reads, not writes. I decided to use the free Data Checker for Peachtree tool for measuring read times because what it does is simply reads all the records in the Peachtree database to see if they are ok; there is nothing else that it is doing. By running the Data Checker in different configurations and timing how long before the report appears I was able to get a single performance number that was very consistent. Repeating the measurement several times gave me results that had an uncertainty of 0.1 second.
The first test was to run the Data Checker against the 2006 Bellwether Garden Supply sample data. When Btrieve 6.15 was used, the processing time was 5.5 seconds. When Btrieve 9.1 was used, the processing time was 5.4 seconds. A slight improvement, not surprising since some improvement is to be expected over the years, but not enough of an improvement to be really significant to users. A few percent faster reads would hardly be noticable to users since there are other factors that cause more variation in performance.
The second test I ran was to compare 2007 Premium with Quantum. The numbers were 6.0 seconds and 5.4 seconds. So Quantum appears to be about 10% faster because data compression is turned off. However, the tradeoff is that the database files have become much bigger. Peachtree backups are typically about 3 times larger and the JrnlHdr.dat file is typically about 8 times larger. Again it is not clear that this slight improvement is of any significance. I doubt that many people would even notice 10% improvement. You can judge for your self whether this improvement is worth having your database grow so much larger.
For really large databases the increased size with Quantum creates an additional problem. Since the JrnlHdr.dat file increases by about 8-fold, if you previously had a JrnlHdr.dat file that was 250 MB or larger then with Quantum that would increase to over 2 GB. Unfortunately the DynaZIP functionality that Peachtree uses to create the backup files has a limit of 2 GB on files it will compress, so you will no longer be able to back up your database from within Peachtree, which means you can't run Data Verification or close the accounting year, both of which force a backup first.
You may wonder why turning off data compression for Quantum doesn't lead to more improvement in reading time. The answer is that the Btrieve compression algorithm is very efficient and compression/decompression is actually a small part of what Btrieve has to do when it processes read requests. Also uncompressed records are larger, so there may be some additional disk overhead in reading them.
The reason I looked at read timing and not write timing is because the vast majority of operations that are performed by Peachtree are reads, not writes. I decided to use the free Data Checker for Peachtree tool for measuring read times because what it does is simply reads all the records in the Peachtree database to see if they are ok; there is nothing else that it is doing. By running the Data Checker in different configurations and timing how long before the report appears I was able to get a single performance number that was very consistent. Repeating the measurement several times gave me results that had an uncertainty of 0.1 second.
The first test was to run the Data Checker against the 2006 Bellwether Garden Supply sample data. When Btrieve 6.15 was used, the processing time was 5.5 seconds. When Btrieve 9.1 was used, the processing time was 5.4 seconds. A slight improvement, not surprising since some improvement is to be expected over the years, but not enough of an improvement to be really significant to users. A few percent faster reads would hardly be noticable to users since there are other factors that cause more variation in performance.
The second test I ran was to compare 2007 Premium with Quantum. The numbers were 6.0 seconds and 5.4 seconds. So Quantum appears to be about 10% faster because data compression is turned off. However, the tradeoff is that the database files have become much bigger. Peachtree backups are typically about 3 times larger and the JrnlHdr.dat file is typically about 8 times larger. Again it is not clear that this slight improvement is of any significance. I doubt that many people would even notice 10% improvement. You can judge for your self whether this improvement is worth having your database grow so much larger.
For really large databases the increased size with Quantum creates an additional problem. Since the JrnlHdr.dat file increases by about 8-fold, if you previously had a JrnlHdr.dat file that was 250 MB or larger then with Quantum that would increase to over 2 GB. Unfortunately the DynaZIP functionality that Peachtree uses to create the backup files has a limit of 2 GB on files it will compress, so you will no longer be able to back up your database from within Peachtree, which means you can't run Data Verification or close the accounting year, both of which force a backup first.
You may wonder why turning off data compression for Quantum doesn't lead to more improvement in reading time. The answer is that the Btrieve compression algorithm is very efficient and compression/decompression is actually a small part of what Btrieve has to do when it processes read requests. Also uncompressed records are larger, so there may be some additional disk overhead in reading them.
