• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Silly power requirements, an over hyped myth?

Even if you buy a 2000w psu, surely it's not drawing 2000w the whole time, and if it's not drawn by components, it won't be an energy hog. I think at present wattage = more connectors, which is probably why more people buy these big units for sli etc. So why worry so much about having a penis enlarger inside your case, it won't be doing your electricity bill any more damage. I honestly can't see the point in spending over £100 on a PSU though.

I'm still waiting for triple/quad sli/crossfire which is the point when i actually invest in a decent PC and PSU again, along with an 8 core cpu (times 2, 16 cores hehe).

Matthew
 
About the 8800 2d/3d, I read somewhere that when you boot, before/while your drivers load, the card defaults a really slow clock speed.

With Rivatuner you can create clock or fan speed profiles which launch on temp thresholds or when Hardware Acceleration is needed. So you run at say, half clock speeds and 60% fan when idling and it auto cranks up to full clocks and 90% fan when you launch a game or when the temp hits a threshold.

I've not played with my 8800 clock speeds yet but having the fan change speed automatically on Hardware acceleration=1 is very cool.

cells why didn't you use speedstep/c1e for the downclocking of your C2D? Does RMclock have to be open all the time?
 
Tom, if you're still up for running tests could you say what the peak power use is while running the power supply test that is part of s&m (some russian tool which is good for stressing components, available here: http://www.benchmarkhq.ru/be_cpu.html).

Oh and while I agree that using 1kw psu's is overkill right now, there are some benefits. One is that the psu will be running at only 40-60% of its rated output and thus its components will be far less stressed (I would worry about running a psu at 85+% of its rated output 24/7 for any great length of time). Another is that some of these new 1kw units are actually some of the most well-built ones around (I'm thinking of the thermaltake toughpower 1000w/1200w, review here: http://www.jonnyguru.com/review_details.php?id=115). Their efficiency puts other brands to shame and they remain rock-solid in terms of voltage regulation and ripple.

Yes they cost an arm and a leg, but when most of the guys using them are already splashing out oodles of cash getting sli'd 8800ultras/crossfired 2900xts, water/phase-cooling for their rig, 24+" monitors, quad-core extreme editions, I doubt they are going to be squeamish about spending in the region of £175-250 for a power supply.
 
Last edited:
Nullvoid said:
Tom, if you're still up for running tests could you say what the peak power use is while running the power supply test that is part of s&m (some russian tool which is good for stressing components, available here: http://www.benchmarkhq.ru/be_cpu.html).

Oh and while I agree that using 1kw psu's is overkill right now, there are some benefits. One is that the psu will be running at only 40-60% of its rated output and thus its components will be far less stressed (I would worry about running a psu at 85+% of its rated output 24/7 for any great length of time). Another is that some of these new 1kw units are actually some of the most well-built ones around (I'm thinking of the thermaltake toughpower 1000w/1200w, review here: http://www.jonnyguru.com/review_details.php?id=115). Their efficiency puts other brands to shame and they remain rock-solid in terms of voltage regulation and ripple.

Yes they cost an arm and a leg, but when most of the guys using them are already splashing out oodles of cash getting sli'd 8800ultras/crossfired 2900xts, water/phase-cooling for their rig, 24+" monitors, quad-core extreme editions, I doubt they are going to be squeamish about spending in the region of £175-250 for a power supply.

No problem, I still have the unit and will give that a go later.
 
Running an 8800GTX on a 460Watt Zalman PSU. Its all about the amps. Mine has 34V for the combined 12V. My 8800GTX only draws 28a according to EVGA and it works perfectly.

If you have a decent PSU then the 12V figures are what you need to concern yourself with.
 
Nullvoid said:
Oh and while I agree that using 1kw psu's is overkill right now, there are some benefits. One is that the psu will be running at only 40-60% of its rated output and thus its components will be far less stressed (I would worry about running a psu at 85+% of its rated output 24/7 for any great length of time).
Trouble is SMPS phased switching PSU are often inefficient with very light loads. Most have a peak efficient point, that'll be the one they quote btw. The trick is to balance loading against the maximum point and avoid running too near the upper limit as that's not good either. The upshot of running an uber PSU with a light load is that it could well be chucking out more heat and wasting more energy than than a smaller PSU loaded correctly.

 
AWPC said:
Running an 8800GTX on a 460Watt Zalman PSU. Its all about the amps. Mine has 34V for the combined 12V. My 8800GTX only draws 28a according to EVGA and it works perfectly.

If you have a decent PSU then the 12V figures are what you need to concern yourself with.
For example a Zalman ZM460-APS has a combined 12v output of 360W at 12v or 30A. And a 8800GTX at stock uses something like 130W at 12v or approx 11A.
 
Would be really interesting to see how much power pulls of a system that consists of, 2 GTXs, a Quad Core CPU, 4gb Of Memory and say 4 Hard Drives?

Then take out the 2 GTXs and just swap them for 2 HD2900XTs.

Come on someone must have a system like that? :p
 
fornowagain said:
Trouble is SMPS phased switching PSU are often inefficient with very light loads. Most have a peak efficient point, that'll be the one they quote btw. The trick is to balance loading against the maximum point and avoid running too near the upper limit as that's not good either. The upshot of running an uber PSU with a light load is that it could well be chucking out more heat and wasting more energy than than a smaller PSU loaded correctly.


I think if you're buying a toughpower 1kw when your system will be running at loads of between 65w-200w then you need shot :)
 
Is the wattage on a PSU what it can handle from the mains or what it can push out once the AC it draws is turned into DC? I think it's the latter. Someone correct me if I'm wrong, but putting a wattage meter on the mains won't tell you what the PSU is putting out in DC form... :confused:
 
Nullvoid said:
I think if you're buying a toughpower 1kw when your system will be running at loads of between 65w-200w then you need shot :)
The curves are an example only. Each one is different and you would need to see the efficiency curves for each PSU before you can make any decisions based on load. In that example a 200/250W load is more efficient on the 300W PSU than the 600W. Also power saving features if enabled, can put idle loads quite a way from peak values.

For instance, the peak efficiency for a 1kW PSU is more likely near the 800W mark and could fall away below 450w. And a 700W psu may peak around 550W and fall away around 300W. Most are bell curves and not flat, but flatter the better and a sign of quality. My point is (other than the number/type of connections and regulation which is another issue) if you care about ongoing cost and heat. Then you should at least consider efficiency and matching PSU's to load and not just buy a big PSU for spare capacity that may never be used.
 
Last edited:
Popular auction site
icon14.gif
 
fornowagain said:
The curves are an example only. Each one is different and you would need to see the efficiency curves for each PSU before you can make any decisions based on load. In that example a 200/250W load is more efficient on the 300W PSU than the 600W. Also power saving features if enabled, can put idle loads quite a way from peak values.

For instance, the peak efficiency for a 1kW PSU is more likely near the 800W mark and could fall away below 450w. And a 700W psu may peak around 550W and fall away around 300W. Most are bell curves and not flat, but flatter the better and a sign of quality. My point is (other than the number/type of connections and regulation which is another issue) if you care about ongoing cost and heat. Then you should at least consider efficiency and matching PSU's to load and not just buy a big PSU for spare capacity that may never be used.

I did get your point the first time around, I just think that for some of these new breed of 1kw supplies, the only big negative is that of price. Referring back to this review you'll see the efficiency numbers are above 80% at the important 20% (well 23...), 50% and 100% loads. It does dip below 80% when it is crossloaded, but I think no modern system will have such a huge load on the 3.3v and 5v rails with next to no load on the 12v rails. As such the unit itself doesn't strictly have a peak efficiency as it is hitting the same high numbers across the majority of its range.

I think we're reaching a point where any decent modern psu will guarantee good efficiencies/voltage regulation+ripple, which leaves us with only price/modularity/power needs/warranty/noise level as the criteria when choosing which supply to buy.
 
Last edited:
Nullvoid said:
I did get your point the first time around, I just think that for some of these new breed of 1kw supplies, the only big negative is that of price. Referring back to this review you'll see the efficiency numbers are above 80% at the important 20% (well 23...), 50% and 100% loads. It does dip below 80% when it is crossloaded, but I think no modern system will have such a huge load on the 3.3v and 5v rails with next to no load on the 12v rails. As such the unit itself doesn't strictly have a peak efficiency as it is hitting the same high numbers across the majority of its range.

I think we're reaching a point where any decent modern psu will guarantee good efficiencies/voltage regulation+ripple, which leaves us with only price/modularity/power needs/warranty/noise level as the criteria when choosing which supply to buy.

Not all pulse switching psu are the same, some of the better ones use feedback circuits to maintain efficiency. As I pointed out a flat curve is desirable and is not the same for every manufacturer. I'm well aware of the larger Toughpowers (good review on [H]), all have good efficiency curves. Again, as I said its a mark of quality and be should considered as a factor in any purchase. Don't take it for granted, some of the PCP&C have terrible efficiency even though they cost a fortune.

http://enthusiast.hardocp.com/article.html?art=MTMzMCw4LCxoZW50aHVzaWFzdA==
This leads us to the part of the Thermaltake advertising that has tripped up most companies, efficiency. In various places the Toughpower is billed as being anywhere from 80% to 87%+ efficient. The low end of this range is almost what we would call expected in this era with power supplies as manufacturers have been pressing units through the 80 Plus program, but 87% is surpassing the numbers associated with the longtime "king" of efficient power supplies Seasonic. With this in mind we approached the Toughpower with a skeptical eye as surely this was only marketing, right? Wrong. At no point during our testing did the Toughpower even bottom out at 80%. The lowest efficiency numbers we could get out of the unit were ~82% at a 1199w load at 100v input. The unit peaked at 87.5% efficiency during the 120v testing when running a load of 609w. These numbers for a SMPS are just amazing. Until now a consumer power supply being 87% efficient was unheard of.
 
Last edited:
Back
Top Bottom