AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2017.06.27 20:51 "[Tiff] Excessive memory allocation while chopping strips", by Nicolas RUFF
2017.06.27 22:08 "Re: [Tiff] Excessive memory allocation while chopping strips", by Even Rouault
2017.06.27 22:43 "Re: [Tiff] Excessive memory allocation while chopping strips", by Olivier Paquet
2017.06.27 22:52 "Re: [Tiff] Excessive memory allocation while chopping strips", by Even Rouault
2017.06.28 02:26 "Re: [Tiff] Excessive memory allocation while chopping strips", by Olivier Paquet
2017.06.28 10:09 "Re: [Tiff] Excessive memory allocation while chopping strips", by Even Rouault
2017.06.28 13:49 "Re: [Tiff] Excessive memory allocation while chopping strips", by Olivier Paquet
2017.06.28 14:36 "Re: [Tiff] Excessive memory allocation while chopping strips", by Even Rouault
2017.06.28 15:22 "Re: [Tiff] Excessive memory allocation while chopping strips", by Olivier Paquet
2017.06.28 15:42 "Re: [Tiff] Excessive memory allocation while chopping strips", by Bob Friesenhahn
2017.06.28 19:01 "Re: [Tiff] Excessive memory allocation while chopping strips", by Rob Tillaart
2017.06.28 20:37 "Re: [Tiff] Excessive memory allocation while chopping strips", by Bob Friesenhahn
2017.06.29 05:52 "Re: [Tiff] Excessive memory allocation while chopping strips", by Paavo Helde
2017.06.28 18:08 "Re: [Tiff] Excessive memory allocation while chopping strips", by Paavo Helde

2017.06.28 20:37 "Re: [Tiff] Excessive memory allocation while chopping strips", by Bob Friesenhahn

wrt the question:

Very large strips and or tiles are used by people who expect that it will increase the performance of their application in some sense, and yes there might be some gain.Most computers can handle MB easily (even

Even now, most (Intel) CPUs have L2 caches optimized for looping through no more than 256k data at a time. Even if more data is read into RAM, the processing loops should usually not consume more than about 128k and performance definitely drops off once the working set exceeds 256k.

The law of diminishing returns is always present, and given L1, L2, and (possibly) L3 caches, there is a diminishment whever there are accesses to the next level of cache, or main memory.

P.S. I live in Texas, where bigger is always assumed to be better.

Bob
--
Bob Friesenhahn
bfriesen@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/