Doom9's Forum - View Single Post

Khun_Doug · 19th March 2018, 16:42

I think the PreFetch tool is the key to what you want. Some filters get no benefit to using PreFetch, while others get drastic improvement. I believe the answer is that StaxRip is always in MT mode, but the PreFetch command is necessary to determine how the MT is invoked. I found it necessary to use PreFetch with mClean.

Depending on your source content, PreFetch (6) may not be sufficient. I am using an i7-6850K, 12 threads. With SD content, the PreFetch (6) is good. But for HD content I needed to go to PreFetch (8) and sometime PreFetch (10). YMMV, but experiment.

Lastly, if you are using filters that require PreFetch and you are doing two or three pass encodes, I highly recommend using the option to save as lossless AVI. I have a few encodes that were three pass X265 with mClean and PreFetch(6). Without the lossless AVI option, I would be running that filter across the same content three times! This way, it is one pass and then the real encoding can begin without sucking up the CPU unnecessarily.

19th March 2018, 16:42	#2872 \| Link
Khun_Doug Registered User Join Date: Jul 2016 Posts: 100	I think the PreFetch tool is the key to what you want. Some filters get no benefit to using PreFetch, while others get drastic improvement. I believe the answer is that StaxRip is always in MT mode, but the PreFetch command is necessary to determine how the MT is invoked. I found it necessary to use PreFetch with mClean. Depending on your source content, PreFetch (6) may not be sufficient. I am using an i7-6850K, 12 threads. With SD content, the PreFetch (6) is good. But for HD content I needed to go to PreFetch (8) and sometime PreFetch (10). YMMV, but experiment. Lastly, if you are using filters that require PreFetch and you are doing two or three pass encodes, I highly recommend using the option to save as lossless AVI. I have a few encodes that were three pass X265 with mClean and PreFetch(6). Without the lossless AVI option, I would be running that filter across the same content three times! This way, it is one pass and then the real encoding can begin without sucking up the CPU unnecessarily.