Haswell to require new Power Supply?

[moonbogg]

lol? If true then you can count me out for sure. This, ontop of weak IPC improvements would stear me well clear of Haswell. I can feel myself getting tired just thinking about Haswell. I think I need a nap.

http://www.guru3d.com/news_story/intel_haswell_low_power_states_could_require_a_new_psu.html

 

[Rakehellion]

Only if you want the new low power states. Basically, you aren't missing out on anything if you don't have a power supply that can't handle it.

 

[Revolution 11]

The real problem is that PSU makers don't report their min amps on the spec sheet. You have no way of knowing which PSU is fine and which is not on launch day.

 

[2is]

Ill bet just my start up processes running would be enough load to keep the CPU from ever entering C6/C7.

 

[frozentundra123456]

Only if you want the new low power states. Basically, you aren't missing out on anything if you don't have a power supply that can't handle it.

Well, the article did say you could have stability issues if you entered the low power state with a PSU that was not designed for it. Seems like there is a setting in the bios to keep from going into that state though.

 

[ShintaiDK]

The original story at vr-zone is completely wrong.

 

[Bill Brasky]

Really these c states are aimed at laptops, which will have compatible power supplies anyway. In every thread about haswell, it seems like there are desktop users screaming about how they don't care about power consumption. But now that it requires a new PSU, we care a lot.

 

[gevorg]

I don't see anything wrong with this requirement. Why should optional features of a new CPU be restricted to 5+ year old standards (ATX12V v2.3)?

 

[MisterMac]

Don't you feel really stupid for posting this drama [redacted] moonbogg?

In the area they're aimed at - they'll meet those specs spot on.

Could you name me a scenario where a self build (any oem WILL use a new PSU with required specs) - would want these low power features?

Any laptop\mobile product would automaticly support it.

I don't get it.



No personal attacks, please.

Also, please avoid profanity, even mild ones, in the tech forums.

Moderator jvroig

 

[nitromullet]

What a sensationalist thread title for a complete non-issue.

 

[blackened23]

Don't you feel really stupid for posting this drama shit moonbogg?

In the area they're aimed at - they'll meet those specs spot on.

Could you name me a scenario where a self build (any oem WILL use a new PSU with required specs) - would want these low power features?

Any laptop\mobile product would automaticly support it.

I don't get it.

From his link:

Older or bargain basement power supply units may be unable to achieve this, and might become unstable or trigger a UVP/OVP shutdown once the system enters the C6/C7 power states. This is because the previous ATX12V v2.3 design guidelines for Ivy Bridge and its predecessors only called for a minimum load of 0.5 Amps on the CPU power rail, hence a less sophisticated internal feedback loop/protection could be used in older or cost-reduced offerings, which will not support the new standards.

I believe that's why he posted it. But, it's not like all existing PSUs won't work, nearly all will. I think the bargain basement 10 year old PSUs may have issues, but who would use one of those anyway.

 

[Abwx]

Switching mode PSUs of all kind need in principle to be minimaly
loaded to damp the output such that the feedback is inherently
slowed and voltage regulation wont display second order behaviour ,
read oscillations converging to the final value once there s a brutal
change in output voltage due to current abruptly rising.

In principle most if not all PSUs should be stable even when not
externaly loaded since there is internal resistors that are used
to drain a slight current to prevent such behaviours , not counting
that the PSU fan is systematicaly connected to the +12V rail ,
providing an added loading.

All in all , i m curious to see if this issue is a real one with members
that hang by there since it will be telling about the care in designing PSUs.

 

[Gillbot]

What a sensationalist thread title for a complete non-issue.

Not only that, but why spend hundreds on a CPU, board and other upgrades only to whine over $50 or whatever on a power supply.

 

[BTRY B 529th FA BN]

I just read about this over on JonnyGuru. Found a nice post in the thread too. check it

Post 3 seems to be funny.. read the last line

 

[mindless1]

... But, it's not like all existing PSUs won't work, nearly all will. I think the bargain basement 10 year old PSUs may have issues, but who would use one of those anyway.

... the average non-techy person whose case came with a PSU in it.

... or anyone who just wants a cost effective 500W PSU to run their 120W system and feels like they're already overpaying for more PSU than they need just to avoid the utter-crap-tier of PSUs.

Re: internal load resistors, many PSU can run unloaded due to them but have cross-load issues because those resistors are often a higher ohm value than ideal to keep heat adjacent to (if not actually touching) the capacitors lower.

 

[Headfoot]

Looks like you can just hit a BIOS toggle to get it to work fine even with a cheap PSU

 

[nenforcer]

Looks like you can just hit a BIOS toggle to get it to work fine even with a cheap PSU

Yeah I'm guessing you can disable the C6 / C7 Power step somewhere in the 80 series chipset BIOS if you have a CompUSA Mad Dog Power Supply in you new Haswell build.

 

[VirtualLarry]

if you have a CompUSA Mad Dog Power Supply in you new Haswell build.

LOL!

 

[UaVaj]

this is one of those stupid green decision. (sounds like ethanol for pc)

given the amout of power saved from haswell vs the the amount of power required to built the new psu and the amount of power required to dispose the old psu.

haswell = fail.

 

[Idontcare]

this is one of those stupid green decision. (sounds like ethanol for pc)

given the amout of power saved from haswell vs the the amount of power required to built the new psu and the amount of power required to dispose the old psu.

haswell = fail.

That's a short-sighted perspective in my opinion.

To get Haswell you are already accepting the idea that you are replacing your existing mobo and existing CPU (or buying a new system entirely).

Is it really all that different from say buying AM3+ and needing DDR3 ram to replace your DDR2 ram if upgrading from AM2 platform?

Here you are buying parts for an upgrade, one of those parts can include a new fancier PSU (bronze to gold or platinum?) that when combined with Haswell enables even better performance/watt.

To me your argument "Haswell = fail" is contrived, over-reaching, and reflective an arbitrariness on your part that stems from an anti-Intel objective.

 

[A5]

this is one of those stupid green decision. (sounds like ethanol for pc)

given the amout of power saved from haswell vs the the amount of power required to built the new psu and the amount of power required to dispose the old psu.

haswell = fail.

You can turn it off in 30 seconds in the BIOS if you have a bad PSU that isn't compatible.

And it's a feature targeted at laptops anyway.

 

[Xpage]

I enjoy sleep to boot on a SSD, waiting a few seconds for a cold boot if your PS doesn't support a sleep state in a minor issue, especially with UEFI and a SSD in your system like most AT users have

 

[wsaenotsock]

CompUSA Mad Dog Power Supply in you new Haswell build.

These PSU's are especially useful for making sure you upgrade at regular intervals. Once they short out all of your components, it removes that worrying question of whether the latest generation chip is a worthy upgrade and helps you pull the trigger. Highly recommended

 

[UaVaj]

These PSU's are especially useful for making sure you upgrade at regular intervals. Once they short out all of your components, it removes that worrying question of whether the latest generation chip is a worthy upgrade and helps you pull the trigger. Highly recommended

now that is a real reason to get haswell.

 

[UaVaj]

That's a short-sighted perspective in my opinion.

To get Haswell you are already accepting the idea that you are replacing your existing mobo and existing CPU (or buying a new system entirely).

Is it really all that different from say buying AM3+ and needing DDR3 ram to replace your DDR2 ram if upgrading from AM2 platform?

Here you are buying parts for an upgrade, one of those parts can include a new fancier PSU (bronze to gold or platinum?) that when combined with Haswell enables even better performance/watt.

To me your argument "Haswell = fail" is contrived, over-reaching, and reflective an arbitrariness on your part that stems from an anti-Intel objective.

no anti intel.

just the nastly truth no body want to hear or acknowledge. afterall haswell (or ethanol) is suppose to be green!