Op woensdag 06 februari 2002 13:29 schreef Four het volgende:
Ben bezig met de specefieke instellingen van mijn bios maar heb nog een aantal vraagjes over de betekenis van de verschillende componenten. En waar kan ik beter zijn dan Forum Nr.7?
Goed wat zijn de volgende dingen
*Spread Spectrum
*(onderdeel) Skew Ajust
*Bank Interleave
En hoe rekent de command rate? Wat doet het waar gaat het van uit?
bank interleave:
quote uit search
DRAM bank interleaving enables the system to alternate memory read/writes between different pages of RAM. This means that it can write some data to one page and then immediately write the next data to another page, without having to wait for the first write to finish. The SDRAM architecture allows for a single RAM chip (or bank) to be split into two pages allowing two-way interleaving within a single chip. As a single 128MB stick of RAM occupies two memory banks, then this allows four way interleaving - hence the observed improvements with 4-way interleaving enabled with only one stick of RAM. With a single 64MB stick of RAM (that occupies just one bank) then one would only see improvements with 2-way interleaving enabled.
of
Bankinterleaving
Spread spectrum:
Spread Spectrum Modulation was invented to reduce interferences of high order harmonics of the bus frequency. The theory is that, because every wave form generates higher order harmonic waves or Obertoene, accumulation of the latter can result in interference with the original signal. One way to avoid this problem is to subject the base frequency to a slow (ca 100,000 clock cycles) modulation, meaning that the FSB varies between e.g. + and - 1% of the nominal value. In older boards, usually two different settings were available, either centered around the nominal value or set with the nominal frequency as the maximum (low modulation). Most current boards employ the centered modulation.
This is, at least, the official version of Spread Spectrum Modulation. In reality, there are different reasons for its implementation. With increasing operating frequency, electronic components emit electromagnetic interference signals (EMI). EMI, on the other hand can cause interferences with other devices and is, therefore, subject to regulation by the FCC which limits the signal amplitude according to its guidelines. Any device exceeding the maximum allowable signal strength will not gain approval by the FCC and can, therefore, not be marketed. In order to understand the reason for SSM, it is necessary to know how the FCC tests EMI. Basically, the testing device is a radio receiver and the testing is done by sweeping its receiving frequency through the frequency range of interest and measuring the interference with the video and audio signals. The bandwidth sensitivity of the measuring device is in the order of about 1 MHz.
If the operating frequency is modulated to spread over a bandwidth of typically 4-5 MHz, the same will happen to the EMI spectrum and, instead of showing a sharp peak, the spectrum will be spread to a more or less Gaussian bell shape. In this case, the amplitude will be, of course, substantially smaller, that is, in the order of 1/3 - Πof the original peak. The energy, however, will be the same. On the other hand, the measuring instrument with its bandwidth limited to only 1/4 of the spread will, consequently, only see 1/3 to Πof the EMI. Therefore, the system will obtain FCC approval even if it exceeds the guidelines.
Recommended Settings
If running at stock setting, enabling Spread Spectrum Modulation (SSM) may reduce EMI caused by accumulation of high frequency components and, thus, add stability, even though I have never been able to confirm this. Under overclocking conditions, enabling SSM may cause a system to crash, simply because with the high multiplier values employed now, even a 0.5% modulation up and down can cause differences up to as much as 10 MHz clock speed within one modulation cycle. In other words, if the CPU is already operating at its limit, increasing the clock speed by another 10 MHz may be fatal. Therefore, for any overclocking, SSM should be turned off.
Another side effect of SSM is that it can interfere with the clock generator. This means that, instead of merely initializing SSM, it is possible to enable FSB settings that were never supported by the manufacturer. Examples are the Tyan Trinity 400 where disabling SSE results in an actual bus speed of 90 MHz instead of the selected 117 MHz or the MSI 6309 where FSB settings of up to 200 MHz become available. The reason is that activating SSM can cancel out the FSB setting since there can be a pin address overlap on the clock generator chip.
Skew adjust: geen idee
/u/38899/lainavantar3.JPG?f=community)
:strip_icc():strip_exif()/u/29357/ninja.jpg?f=community)