Jump to content

moloko

Quark
  • Content Count

    6
  • Joined

  • Last visited

Community Reputation

0 Serf

About moloko

  • Rank
    Serf
  1. Good point to raise as PSU manufacturers do change OEMs as new models are released within a range. So for clarification, the higher wattage units, i.e. 850W, 1000W, 1200W units from the DA and OP ranges are based on the Seventeam OEM design - that is unless they've been updated, heh. Also worth mentioning at this point that while a manufacturer may use an OEM design, it does not mean that the manufacturer's model and the equivalent OEM model is the same. This is because the manufacturers make changes to everything from the length of the cables through to the choice of capacitors to use. One such example is the Corsair HX1000 has specially chosen capacitors (unfortunately father time has removed the rest of the specifics from my memory from my last trip to Corsair HQ - soz) Fair enough that the list, erm, lists, the number of rails. This is fine for those that are informed by the differences and influences of single rail vs multiple rail PSUs but sometimes people simply need to be told what's what so they can at least get their foot in the door for a good future choice. My suggestion of a list of 'above average' PSUs lends itself to both giving an answer to the utterly uninformed, giving props to the vendors that are that cut above the rest and also for those wanting the unquestionably famed PSU. For example, my background for PSU choice comes down to being a benching enthusiast therefore I have an inclanation towards high amperage single rail solutions. If someone was looking to buy some kit to get into the benching scene there is a specific list of the 'it' PSUs for bench systems such as the Corsair TX750, Silverstone OP1000, Silverstone DA1200, PC Power & Cooling Silencer 750W etc Simple answer: Yes Extended answer: I wouldn't use PSUs from the likes of Zalman (known to replace internal parts with cheaper alternatives; feel free to draw the conclusion on this one), a number of the Enermax units (specifically Infinity series), low wattage multiple rail units such as the Antec EarthWatch are also a unit to stay clear of (30A over 2 rails? No thanks) and even some of the larger wattage multiple rail units such as the Coolermaster Real Power Pro 1250W that has from memory 6 rails with half running 28A and the other half running 20A; hardly a beast compared to the single rail OP1000 and DA1200. But this is floating back towards a preference of both quality of parts within the unit and the quality in the design of the unit. In my day to day benching I use a Corsair TX750 and Silverstone OP1000 and have the Thermaltake ToughPower 1200W sit on the shelf as a backup because it has four rails rated for 20A on the two smaller rails and 36A on the two larger rails. The Tt unit only comes out when I'm benching extremely high frequencies and use the extra rails on the GPUs with a single unit supplying the ATX and EPS power as well as the HDD and fans - the next range of ToughPower units will be single rail.
  2. moloko

    What Sort Of Overclock Are You Running

    For my 24/7 work PC I run an E6700 at a relaxed 3.2GHz (it's my trusty old 4.1GHz air chip; over 5GHz under cascade) cooled by a Noctua NH-U12F. On a daily basis I'm running chips anyware from 4GHz to 6+GHz depending on the chip and cooling (air, water, phase or LN2) that I'm using at the time.
  3. moloko

    CPU influences on pi result

    Depends on the calculation being done. As the calculation gets bigger it gets more and more dependent on the system memory as it moves out of the limits of the CPU cache (i.e. 1M to 32M). For example now days most the 1M calculation is done in the CPU cache hence why higher cache 45nm Core 2 Duos have a significant advantage over the 65nm Core 2 Duos and why both decimate Athlon64 and Phenom architecture (not to mention Core is just more efficient but that a discussion for another thread entirely). Here's a basic checklist of things that can and will impact upon a Super Pi time: CPU frequency - higher is better FSB frequency - higher is better Memory frequency - higher is better Memory latency - lower is better MCH latency - lower is better That's the core basis of obtaining a quicker Super Pi time. Another very important aspect to Super Pi benching is the efficiency at which the system operates at these frequencies. The efficiency is largely dictated by the allocation of system resources and the priority and availability of accessing these resources to obtain higher efficiency. Some areas of interest when looking to improve system efficiency: Cache optimisation both in regards to CPU cache and system cache; research the Copy Waza tweak for one such example and others such as LargeSystemCache and CD-T System priority and operation priority can be taken care of using task manager to set the process/task priority and also allocate to a single core on a multi-core processor to further optimisation and not spend resources where it is not necessary i.e. switching a process from one core to another Restricting the maximum memory addressable within the system to optimise cache usage (the optimum value changes depending on the size of the calculation being done) This is just the tip of the iceberg, there are unique tips and tricks within a certain tweak and many areas of a tweak are only learnt through practice; can be difficult to translate the understanding in a comprehensible sentence/paragraph - this is why extreme overclocking is considered an art among those who practice it and such respect is given to those who can master it. It's not the scores that make an overclocking master but the understanding of why the things that are happening are indeed happening. Good luck ^_^
  4. moloko

    Atomic Superpi Scores V5

    zeb's result wasn't using SuperPi Mod 1.5 and therefore doesn't apply to the rules of this little results collective. While it's neat to see what a multi-threaded Pi calculation can be done in, it's the single threaded solution that is the prize here.
  5. GIGABYTE GA-X48T-DQ6 Foxconn X48 BlackOps Biostar TPower i45 ASUS Commando ASUS P5B Deluxe While I believe there were plenty of great boards in the past, if we're talking about the best, well, they keep on getting better. With that said, props to the P5B Deluxe and Commando for making the list - still great boards.
  6. FYI, Silverstone OEM is Seventeam. Also a recommendation for a list of 'above average' PSUs; basically a list of top recommendations (tier 1 units) of which would largely be made up of single rail PSUs. Such a list would contain Corsair TX series, Silverstone OP and DA series etc. Keep an eye out for a new range of Thermaltake ToughPower PSUs coming shortly; they should be the biz.
×