Version 2.0.0 overview

Few things good to know:

1. It is recommended to run the program as administrator or with disabled UAC.
   
2. If you don\'t need real time monitoring, leave it disabled, as it need some CPU power and most likely will lower your score.
   
3. Real time monitoring may not work on certain Windows 10 builds (1607).
   
4. Sometimes more than one instances of OHM launch when real time monitoring is selected. In that case please close all but one.
   

Real time monitoring supports two modes:

1. LCC - CPU clock must not exceed predefined frequency during the benchmark run.
   
2. Ambient - before launching the benchmark run, short idle temperature check is performed. Only when defined rules (minimum temp, allowed CPU idle load) are met, the encoding will launch.
   

AMD CPUs and APUs from 10h and 15h family often report much lower temperature than they realy have. The difference between reality and indicated temp. can be as much as 10 - 20°C. This makes temperature verification for ambient competitions difficult.

HWBOT x265 Benchmark v2.0.0 is based on improved x265 encoder, build 2.0.0.24 (compiled by GCC). Compared to previous version (1.7.0.374 GCC), the new one greatly improves encoding speed on virtually all processor architectures. Also multi threading seems to be better - the bigger your CPU is, the better relative fps gain you can expect. Here you can see performacne figures for most common CPU architectures.

Measured performance gain (64-bit, 1080p):

• NetBurst (1C/2T, SSE3) - 17%
  
• Core 2 (2C/2T, SSE4.1) - 28%
  
• Nehalem (4C/8T, SSE4.2) - 35%
  
• Sandy Bridge (4C/4T, AVX) - 34%
  
• Sandy Bridge (8C/16T, AVX) - 41%
  
• Broadwell (8C/16T, AVX2) - 53%
  
• Broadwell (20C/40T, AVX2) - 71%
  
• K8 (2C/2T, SSE2) - 16%
  
• K10 (4C/4T, SSE2) - 16%
  
• 15h (2C/2T, AVX) - 28%
  
• 15h (32C/32T, AVX) - 48%
  

Measured performance gain (64-bit, 4k):

• Nehalem (4C/8T, SSE4.2) - 78%
  
• Sandy Bridge (4C/4T, AVX) - 76%
  
• Sandy Bridge (8C/16T, AVX) - 89%
  
• Broadwell (8C/16T, AVX2) - 105%
  
• Broadwell (20C/40T, AVX2) - 172%
  
• 15h (32C/32T, AVX) - 102%
  

The last thing worth mentioning is possibility to manually select instruction set to be used by the encoder. This function is available in the main menu - "CPU features override". As the name suggests, you can select pretty much any combinations of instructions when using custom mode. Or use one of the presets, which should cover most of currently available CPUs. For regular benchmarking there is no need to adjust this, x265 will autodetect your CPU and use the best instruction set for it. The main purpose if this is experimenting with various settings. Beware that activating higher instruction than you CPU supports in hardware (E.G. AVX2 on Sandy Bridge) won\'t bring more speed and quite possibly will crash the encoder. Perhaps the only case of positive influence might be overclocked "locked" Skylake. It is possible disabling AVX2 and running fast enough might give you more fps than almost default clock with AVX2 enabled... a.k.a. XTU @ 742 :)