Memory settings in BIOS
-
- Posts: 588
- Joined: Wed Nov 06, 2013 10:01 pm
Memory settings in BIOS
Surely a staple of audio computer setup but I had not bothered with this until now.
I am using the ASRock H81M-DGS R2.0 board. It has the most extensive BIOS settings I have ever used. This is a cheap board so it is a good surprise that it has this much adjustability in the BIOS.
Last night I figured it was time to do something with the memory timing.
The board has a useful explanation for each parameter and the range it will allow. I basically set all of them as low as they would go thinking it would not boot with these settings. I compared them to settings I had used on previous installs and they seemed impossibly low.
One interesting thing is that boot takes about 3 minutes now. I surmise the memory is being cleared and all is being re-loaded, which I like. It is a bother but this is tweaky audio. With the old minimal XP/cMP boot would take a similar amount of time.
Well, no one will be surprised when I say this made a very noticeable improvement. For those using the secret player there is a reduction in forwardness and bass has been brought closer to where it should be. Could be the player is more sensitive to memory setting than we would have thought?
Since there are so many settings I will take photographs of the screen sometime today. Whether these setting will have the same effect with different boards is up to the user to discover.
One thing I liked is that with the Asrock you can disable the unused memory rank in ways I have never seen before. Of course, you usually have the simple on/off off setting but this thing goes much further. When you do these setting you can be pretty well assured the system is not going to bother looking there ever again.
I am using the ASRock H81M-DGS R2.0 board. It has the most extensive BIOS settings I have ever used. This is a cheap board so it is a good surprise that it has this much adjustability in the BIOS.
Last night I figured it was time to do something with the memory timing.
The board has a useful explanation for each parameter and the range it will allow. I basically set all of them as low as they would go thinking it would not boot with these settings. I compared them to settings I had used on previous installs and they seemed impossibly low.
One interesting thing is that boot takes about 3 minutes now. I surmise the memory is being cleared and all is being re-loaded, which I like. It is a bother but this is tweaky audio. With the old minimal XP/cMP boot would take a similar amount of time.
Well, no one will be surprised when I say this made a very noticeable improvement. For those using the secret player there is a reduction in forwardness and bass has been brought closer to where it should be. Could be the player is more sensitive to memory setting than we would have thought?
Since there are so many settings I will take photographs of the screen sometime today. Whether these setting will have the same effect with different boards is up to the user to discover.
One thing I liked is that with the Asrock you can disable the unused memory rank in ways I have never seen before. Of course, you usually have the simple on/off off setting but this thing goes much further. When you do these setting you can be pretty well assured the system is not going to bother looking there ever again.
phonograph, amplifiers & speakers
-
- Posts: 588
- Joined: Wed Nov 06, 2013 10:01 pm
Re: Memory settings in BIOS
Should have said three minutes, actually less, to POST not boot.
Also did the extremely tweaky and strange sprinkling of crystals onto the MB and the DAC.
I am hearing a greatly reduced amount of "digital" noise.
One can enjoy listening a low volume now. Funny how the digital noise makes you want to listen at higher volumes, instinctively, in an effort to mask it and to be able to perceive the dynamics this noise reduces.
For the moment it is as if there is none at all. I am sure over time one will hear it again but it has been diminished to the point that I cannot hear it, yet.
I am thinking the memory timing reduced the noise to the point where the crystals are able to actually absorb what is left.
Also did the extremely tweaky and strange sprinkling of crystals onto the MB and the DAC.
I am hearing a greatly reduced amount of "digital" noise.
One can enjoy listening a low volume now. Funny how the digital noise makes you want to listen at higher volumes, instinctively, in an effort to mask it and to be able to perceive the dynamics this noise reduces.
For the moment it is as if there is none at all. I am sure over time one will hear it again but it has been diminished to the point that I cannot hear it, yet.
I am thinking the memory timing reduced the noise to the point where the crystals are able to actually absorb what is left.
phonograph, amplifiers & speakers
Re: Memory settings in BIOS
what timings are u on now?
sd card player, modded soekris dac, class a lifepo4 amp or gb class a/b amp, diy open baffle speakers based on project audio mundorf trio 10's
-
- Posts: 395
- Joined: Mon Aug 11, 2014 4:19 am
Re: Memory settings in BIOS
I had a couple questions too.
What kind of RAM are you using? Part number would be nice to know.
And what's this "magic dust" that you're sprinkling on the PC and DAC?
Randy
What kind of RAM are you using? Part number would be nice to know.
And what's this "magic dust" that you're sprinkling on the PC and DAC?
Randy
MSI H81-P33 MB, Xeon E3-1225 V3, LPS/LIFEPO4 and Astron RS-12A, 240gb SSD music, 2nd SSD for OS, Mod SS PCIE USB card, Server2012 R2 Ess+AO+MQn, Amanero USB to DACEND ES9018 to Schiit Lyr amp to Senn HD 700 headphones
-
- Posts: 588
- Joined: Wed Nov 06, 2013 10:01 pm
Re: Memory settings in BIOS
I went to the office to get the camera and then left it there. There are too many to write down, though I might try.
The ASROCK board tells you how low you can set so that is what I did.
Memory is nothing special: look here http://www.newegg.com/Product/Product.a ... 6820146742 Running at 1.3V.
I got this because it was the lowest latency I could find and I did not complain that it was cheap.
Magic dust are literally quartz crystals. I think it was the lately departed and much missed Stewart Ono at AA who started this. We were sprinkling crystals on the MBs when cMP was the player. I went further and placed them on the DAC this time. Lots of them - not that it covers the boards entirely but almost.
I got mine on EBAY many years ago. Get different colors they have different qualities.
Nigel, I turned off HPET and have been told the player does use it so I turned it back on. Definitely no need for Hyper Threading if you have that enabled.
There is almost nothing ENABLED in my BIOS. One strange thing is the CPU seems to be running hotter - I had turned off most of the power management stuff at the start but looked deeper (these are scattered throughout the BIOS) and turned off even more of them. I would rather deal with an extra warm block of brass than the MB making arbitrary decisions on powering the CPU. What do you think of that?
I am not hearing anything that would say this heat is a sonic problem. I guess my board is basically staying at 1.00 volt, I wish I could get it lower. There are some adjustments I am not sure about that might allow that. Hate the thought of, after waiting for that long POST, the thing not turning on at all and having to re-set the BIOS so that will come later.
I told Jonathan that the sound is so good I found myself walking softly as to not upset the turntable last night.
The ASROCK board tells you how low you can set so that is what I did.
Memory is nothing special: look here http://www.newegg.com/Product/Product.a ... 6820146742 Running at 1.3V.
I got this because it was the lowest latency I could find and I did not complain that it was cheap.
Magic dust are literally quartz crystals. I think it was the lately departed and much missed Stewart Ono at AA who started this. We were sprinkling crystals on the MBs when cMP was the player. I went further and placed them on the DAC this time. Lots of them - not that it covers the boards entirely but almost.
I got mine on EBAY many years ago. Get different colors they have different qualities.
Nigel, I turned off HPET and have been told the player does use it so I turned it back on. Definitely no need for Hyper Threading if you have that enabled.
There is almost nothing ENABLED in my BIOS. One strange thing is the CPU seems to be running hotter - I had turned off most of the power management stuff at the start but looked deeper (these are scattered throughout the BIOS) and turned off even more of them. I would rather deal with an extra warm block of brass than the MB making arbitrary decisions on powering the CPU. What do you think of that?
I am not hearing anything that would say this heat is a sonic problem. I guess my board is basically staying at 1.00 volt, I wish I could get it lower. There are some adjustments I am not sure about that might allow that. Hate the thought of, after waiting for that long POST, the thing not turning on at all and having to re-set the BIOS so that will come later.
I told Jonathan that the sound is so good I found myself walking softly as to not upset the turntable last night.
phonograph, amplifiers & speakers
Re: Memory settings in BIOS
Jonathan is old enough to remember fiddling with battery powered crystal radio receivers and so are you sure Rick that you are not listening to our excellent Radio 3 on your precisely aligned crystals?
Happy New Year!
Happy New Year!
Maplin XM21X 12V float charging A123 26650 LiFePO4 battery/Maxwell Supercap PSU for Mitac PD10-BI J1900 Bay Trail, WTFPlay, Hiface Evo, Bow Technologies 1704 NOS DAC, StereoKnight TVC, Quad II monoblocks, ZU Audio Druid Mk4/Method Sub
Re: Memory settings in BIOS
Actually I am surprised that lowering the performance of RAM makes an improvement. In an ideal world, surely logic would dictate that we are trying to achieve as close to real time processing of the music file as is possible on the flawed boards and non-real time OS that we are forced to use?
I do not of course doubt that you and others find improvement. However, I suspect that might be because the benefit of the reduction in digital 'noise' obtained by lowering RAM voltage and speed is outweighing the negative effect of slower processing of the audio stream. But what do I know?
I do not of course doubt that you and others find improvement. However, I suspect that might be because the benefit of the reduction in digital 'noise' obtained by lowering RAM voltage and speed is outweighing the negative effect of slower processing of the audio stream. But what do I know?
-
- Posts: 588
- Joined: Wed Nov 06, 2013 10:01 pm
Re: Memory settings in BIOS
The irony is I have sped up the memory but slowed everything else down.
Lowering the timings speeds up the memory.
Those numbers are clock timings that determine how long between the various operations the memory performs.
A setting of 20 means that many ticks of the clock before that aspect is performed. If you set this to 4 it is going to happen five times more often.
So we do not need the CPU to work very hard, there isn't that much work to do but we want the memory to be constantly updated; we do not want stuff to reside in the memory for very long unless it is needed there, like the OS. We want the music files to flow quickly. A terribly simplified explanation but that is the best I can do.
Another thing I found to be important is to not leave any parameter set to AUTO. Don't give the BIOS any decision making. This to me is like the difference in a class A amplifier and a class AB amplifier. There is enough switching going on without adding to it.
Will have some pictures later today.
Lowering the timings speeds up the memory.
Those numbers are clock timings that determine how long between the various operations the memory performs.
A setting of 20 means that many ticks of the clock before that aspect is performed. If you set this to 4 it is going to happen five times more often.
So we do not need the CPU to work very hard, there isn't that much work to do but we want the memory to be constantly updated; we do not want stuff to reside in the memory for very long unless it is needed there, like the OS. We want the music files to flow quickly. A terribly simplified explanation but that is the best I can do.
Another thing I found to be important is to not leave any parameter set to AUTO. Don't give the BIOS any decision making. This to me is like the difference in a class A amplifier and a class AB amplifier. There is enough switching going on without adding to it.
Will have some pictures later today.
phonograph, amplifiers & speakers
Re: Memory settings in BIOS
Thanks for the explanation which is clear and makes perfect (logical) sense.
I had it in my rather unreliable memory that you & nige advocated lowering the speed of your RAM to say 800 MHz or the lowest that one's board's BIOS would permit? In which case my logic would apply.
I had it in my rather unreliable memory that you & nige advocated lowering the speed of your RAM to say 800 MHz or the lowest that one's board's BIOS would permit? In which case my logic would apply.
Re: Memory settings in BIOS
what's the secret player ?rickmcinnis wrote:Surely a staple of audio computer setup but I had not bothered with this until now.
I am using the ASRock H81M-DGS R2.0 board. It has the most extensive BIOS settings I have ever used. This is a cheap board so it is a good surprise that it has this much adjustability in the BIOS.
Last night I figured it was time to do something with the memory timing.
The board has a useful explanation for each parameter and the range it will allow. I basically set all of them as low as they would go thinking it would not boot with these settings. I compared them to settings I had used on previous installs and they seemed impossibly low.
One interesting thing is that boot takes about 3 minutes now. I surmise the memory is being cleared and all is being re-loaded, which I like. It is a bother but this is tweaky audio. With the old minimal XP/cMP boot would take a similar amount of time.
Well, no one will be surprised when I say this made a very noticeable improvement. For those using the secret player there is a reduction in forwardness and bass has been brought closer to where it should be. Could be the player is more sensitive to memory setting than we would have thought?
Since there are so many settings I will take photographs of the screen sometime today. Whether these setting will have the same effect with different boards is up to the user to discover.
One thing I liked is that with the Asrock you can disable the unused memory rank in ways I have never seen before. Of course, you usually have the simple on/off off setting but this thing goes much further. When you do these setting you can be pretty well assured the system is not going to bother looking there ever again.