Romex Software Forum

@krzyys
On file level Windows is thinking the file was written but in real it was not. What is Windows internal doing with such information? I dont know

@support. Thank you for the idea!
If I set the cache to write exclusively, the data in the cache will not be preloaded at boot or login. in older versions that went. this is what i mean.

Good morning,
in older versions, the read and write data wasn´t seperate. If written data are not marked as standby-blocks to free the cache, this data can be prefetch at boot or login. I mean the content itself, whether the data was written or read did not matter.

Thank you for the answer!
No, i mean generaly prefetch of Write-Data to L1 even after login!

Hello Romex software developer! I like to have prefetch written data! Why is this feature gone? What do you think about a dynamic strategy? When the cache is full, it will switch to either write exclusively or buffer from inteligent or something like that!? Sorry for my bad english.. Thank you and h...

Nach den kyrillischen Schriftzeichen zu urteilen, scheint dein Wortschatz, einer slawischen Sprache zu entstammen!
My weet over caching and programming is insufficient.. sorry

Only for speed i think it is better to cache the VHD directly to prevent a doubled transferring. In theory the latency is a little bit smaller by caching one of both things (store or VHD). Informations to my cache: VHD is formated with 4k cluster (NTFS) and its own cache equally. The underlying HDD ...

Hello hawkhawk!
Here a little benchmark! I hope it is usefull.
http://www.directupload.net/file/d/4841 ... q2_jpg.htm

Hi, Alex I can see a doubled rate of multithreading-4k in comparison to my CDM-benchmark! Can you make a test with more threads as 6, like 8 please?! Here my benchmarks with disabled cores. One core: http://www.directupload.net/file/d/4796/zb4f6i46_jpg.htm Two cores: http://www.directupload.net/file...

Hi, Alex Where you disabled the CPU-cores, at device-manager? i like to make benchmarks with disabled cores! Can you make a benchmark with CrystalDiskMark and set sequential-measure to q1 and t6 (Threads can be higher or lower). By your quadchanneling RAMs, the amount of transferrate can be doubled....