AMD not adhering to 125W spec with FX8350?

[anthony566]

Do We care ?? ... I don't not as long as it works !!!!

 

[ubersonic]

What?

It was pretty self explanatory...

Do We care ?? ... I don't not as long as it works !!!!

But it doesn't, that's the whole issue that's arisen, because the FX-8350 runs outside of its own official spec, boards that were designed based on its official spec and not over engineered are incompatible with it and workarounds like the one mentioned in the OP have to be done to avoid problems (that workaround being tricking the CPU into thinking its overheating so it dials down performance)

---------

In case people have become confused by all the cross talk of TDP and power draw in this thread, were not talking about a CPU that's rated at 125w TDP and consumes >125w power, that's a given, we're talking about a CPU that is officially rated and marketed as being 125w TDP yet is actually ≥140w TDP and consumes nearly 200w under full load.

 

[adolf hamster]

where did it say it takes 200w under full load? the op, and indeed the whole thread, is about taking 140w under full load as opposed to 'supposedly' 125w.

edit: my apologies, note to self read all the tabs before replying to a thread....

 

[eddyr]

TDP could be whatever you want it (within reason) regardless of peak power draw through throttling/power managment. Thermal inertia would allow you to use full load/clocks even if unsustainable.
MSI's lack of adequate power delivery seems a seperate issue and without knowing the official specifications given to board manufacturers regarding power requirements discussing with assumptions seems a little pointless.

 

[Hagar]

In case people have become confused by all the cross talk of TDP and power draw in this thread, were not talking about a CPU that's rated at 125w TDP and consumes >125w power, that's a given, we're talking about a CPU that is officially rated and marketed as being 125w TDP yet is actually ≥140w TDP and consumes nearly 200w under full load.

I guess you are talking about my power draw conclusions, but I must emphasize again that was at 4.6GHz on eight cores running prime 95 for maximum power and it is only my spurious assumptions for the rest of system power draw and efficiencies.

I have not run stock comparisons.

I cannot comment on what the TDP is or how motherboard manufacturers have determined it from testing. I will say though that the piledriver seems very similar to the 1090T which has fewer cores at less OC and also that intel i7 sandys with HT on and similar OC can draw similar power as well.

Most informed people would not use a low midrange motherboard on any OC system, with power droop and VRM issues etc.

 

[ubersonic]

I guess you are talking about my power draw conclusions

Ahh no sorry, should have clarified that, the post referenced in the OP is just one of many and another one shows a report done by techreport on the peak draw of many stock CPU's with the FX-3850 maxing out at 196w draw under peak load (more than a stock 3960X and almost double the peak draw of a stock 3770K). This is of course well above what AMD specified the chip would do when giving motherboard manufacturers numbers to work to, thus anyone who didn't over engineer their board has a problem.

 

[CAT-THE-FIFTH]

Ahh no sorry, should have clarified that, the post referenced in the OP is just one of many and another one shows a report done by techreport on the peak draw of many stock CPU's with the FX-3850 maxing out at 196w draw under peak load (more than a stock 3960X and almost double the peak draw of a stock 3770K). This is of course well above what AMD specified the chip would do when giving motherboard manufacturers numbers to work to, thus anyone who didn't over engineer their board has a problem.

Then try running a Core i7 3960X in the following motherboard with a 4+1 phase VRM:

http://www.msi.com/pic/product/five_pictures7_2501_20111215093233.jpg

Edit!!

Even the smallest socket 2011 motherboard from Shuttle does not use a 4 phase VRM either.

Even,then,having said that,how come Biostar seems to use 4 phase VRMs which are fine??

Look at these articles:

http://www.techpowerup.com/reviews/Biostar/TA990FXE/

http://www.anandtech.com/show/5714/...ond-wind-for-asus-gigabyte-msi-and-biostar/15

It seems to be that MSI under designs their motherboards. So,they are shifting the blame to everyone but themselves.

Not surprising that their previous AMD AND Intel motherboards have had VRM failures:

http://www.overclock.net/a/database-of-motherboard-vrm-failure-incidents

 

[ubersonic]

It seems to be that MSI under designs their motherboards. So,they are shifting the blame to everyone but themselves.

So basically AMD said the CPU will use up to X power, but in fact it turns out to use X+Y power, and as MSI only engineered their board with a CPU using X power in mind its somehow their fault not AMD's, because they didn't over engineer their board?

Not surprising that their previous AMD AND Intel motherboards have had VRM failures

A lot of manufacturers have had VRM issues over the years, IIRC Gigabyte had a VRM issue with their early X79 boards that resulted in them releasing a BIOS update to limit performance like MSI's example in the OP.

 

[CAT-THE-FIFTH]

So basically AMD said the CPU will use up to X power, but in fact it turns out to use X+Y power, and as MSI only engineered their board with a CPU using X power in mind its somehow their fault not AMD's, because they didn't over engineer their board?

Nope,its more a case of MSI consistently under-engineering their motherboards as they cost cut repeatedly. I have known two mates who had high end MSI P55 motherboards fail with only a mild overclock on a Core i5 760.

MSI have had much more VRM failures than any brand. Its because they tend to re-use very old VRM designs with many of their motherboards and don't bother to test them properly.

So yes this is the fault of MSI and also,please explain to me the lack of consumer socket 2011 motherboards with 4+1 phase VRMs??

Moreover,it seems you imply yourself that MSI motherboards are marginal,so why should anyone bother with motherboards which cannot even sustain a mild overclock??

However,as you seemed to have missed the previous link,lets post the pictures.

A lot of manufacturers have had VRM issues over the years, IIRC Gigabyte had a VRM issue with their early X79 boards that resulted in them releasing a BIOS update to limit performance like MSI's example in the OP.

So,is that down to Intel giving wrong specifications to Gigabyte or rather Gigabyte making an error somewhere??

However,MSI has the most of them,and moreover their VRMs tend to go pop,rather than even throttle back.

 

[Martini1991]

I've had an MSI VRAM pop, my 1155 board.

There is a difference between having poor VRM's, and having poor VRM's and supporting a CPU that's out of spec.

 

[CAT-THE-FIFTH]

Also,125W is TDP,not power consumption. That is a thermal design parameter,not a power consumption parameter.

AMD has claimed that the FX-8350 to be 125W. However, during our internal testing, when the usage of CPU reaches 100% through Prime 95 for a while, the power consumption can exceed 125W and reaches 140W. With such condition on the 970A-G46, the high amount of power draw also causes the MOSFET to exceed its spec and will overheat.

The chap does not known what power consumption and TDP is:
http://en.wikipedia.org/wiki/Thermal_design_power

You would think,a tech involved in testing and probably development of their motherboards would know the difference. Its not a good sign.

The thermal design power (TDP), sometimes called thermal design point, refers to the maximum amount of power the cooling system in a computer is required to dissipate. The TDP is typically not the most power the chip could ever draw, such as by a power virus, but rather the maximum power that it would draw when running "real applications". This ensures the computer will be able to handle essentially all applications without exceeding its thermal envelope, or requiring a cooling system for the maximum theoretical power (which would cost more but in favor of extra headroom for processing power).
In some cases the TDP has been underestimated such that in real applications (typically strenuous, such as video encoding or games) the CPU has exceeded the TDP. In this case, the CPU will either cause a system failure (a "therm-trip") or throttle its speed down.[1] Most modern CPUs will only cause a therm-trip on a catastrophic cooling failure such as a stuck fan or a loose heatsink.
For example, a laptop's CPU cooling system may be designed for a 20 watt TDP, which means that it can dissipate up to 20 watts of heat without exceeding the maximum junction temperature for the computer chip. It can do this using an active cooling method such as a fan or any of the three passive cooling methods, convection, thermal radiation or conduction. Typically, a combination of methods is used.
Since safety margins and the definition of what constitutes a real application vary among manufacturers, TDP values between different manufacturers cannot be accurately compared. While a processor with a TDP of 100 W will almost certainly use more power at full load than a processor with a 10 W TDP, it may or may not use more power than a processor from a different manufacturer that has a 90 W TDP. Additionally, TDPs are often specified for families of processors, with the low-end models usually using significantly less power than those at the high end of the family.
The dynamic power consumed by a switching circuit is approximately proportional to the square of the voltage:[2]

(where C is capacitance, f is frequency and V is voltage).

http://www.cpu-world.com/Glossary/T/Thermal_Design_Power_(TDP).html

The Thermal Design Power (TDP) is the average maximum power a processor can dissipate while running commercially available software. TDP is primarily used as a guideline for manufacturers of thermal solutions (heatsinks/fans, etc) which tells them how much heat their solution should dissipate. TDP is not the maximum power the CPU may generate - there may be periods of time when the CPU dissipates more power than designed, in which case either the CPU temperature will rise closer to the maximum, or special CPU circuitry will activate and add idle cycles or reduce CPU frequency with the intent of reducing the amount of generated power.
TDP is usually 20% - 30% lower than the CPU maximum power dissipation.

People also forget about batch to batch variation and VIDs.

My old Q6600 was very high VID just like the Q9300 my mate had,which means they were probably on the upper limit of power consumption for those CPUs.

I would expect any motherboard to have some leeway as a result of this,so it does call into question what the heck MSI is doing when they are designing their motherboards.

Edit!!

From the 970-G46 webpage:

http://www.msi.com/product/mb/970A-G46.html

So,they are saying the motherboard is great for overclocking,and yet say it throttles a CPU at stock.

Maybe,in the land of MSI overclocking does not increase power consumption and TDP.

Perhaps,MSI should have just rated the motherboard for a FX6300 instead?? If they rated it for 95W TDP CPUs,then it would mean less issues.

Problem sorted.

 

[Martini1991]

TDP isn't power usage, but power usage should be higher than TDP.
Still out of spec.

 

[humbug]

I have to agree with CAT, i have been completely put off by MSI Motherboards for years now, nothing but problems with them.

That seems to be the case with some of their GPU's now as well.

Thats a shame, as MSI in years gone by were a byword for enthusiasts hardware.

 

[Martini1991]

Yeah, I'm pretty much against MSI stuff myself, I had a 7870 VRM blow as well, only the day after I'd built it, for a mates 21st birthday present.

I don't understand why TDP is ALWAYS put across as power usage.
At any rate, power consumption must be higher than TDP.

 

[ubersonic]

Nope,its more a case of MSI consistently under-engineering their motherboards as they cost cut repeatedly.

Under engineering and engineered to do what is required are not the same thing. Issues arising because MSI didn't over engineer the board are AMD's fault for understating what was required of the board.

So yes this is the fault of MSI and also,please explain to me the lack of consumer socket 2011 motherboards with 4+1 phase VRMs??

Again, no it isn't it's AMD's fault, and what does LGA2011 have to do with anything? :S

So,is that down to Intel giving wrong specifications to Gigabyte or rather Gigabyte making an error somewhere??

GB making an error, the point was MSI are not the only manufacturer to have VRM issues.

Also,125W is TDP,not power consumption. That is a thermal design parameter,not a power consumption parameter.

Correct, as mentioned in this thread by myself and others on a number of occasions.

The chap does not known what power consumption and TDP is

Addressed previously in the thread, its Chinese whispers due to the customer support guy relaying info from R&D.

Perhaps,MSI should have just rated the motherboard for a FX6300 instead?? If they rated it for 95W TDP CPUs,then it would mean less issues.

Problem sorted.

Well if AMD hadn't given out incorrect specs for the FX-8350 to MSI then I'm sure MSI would have either designed the board for an FX-3850 or just rated it for an FX-6300...

 

[CAT-THE-FIFTH]

Under engineering and engineered to do what is required are not the same thing. Issues arising because MSI didn't over engineer the board are AMD's fault for understating what was required of the board.

No,because it appears you are adamant this is the fault of AMD.

This looks more like a classic example,of MSI again underspeccing their motherboards.

They have done it repeatedly for years.

Ahh no sorry, should have clarified that, the post referenced in the OP is just one of many and another one shows a report done by techreport on the peak draw of many stock CPU's with the FX-3850 maxing out at 196w draw under peak load (more than a stock 3960X and almost double the peak draw of a stock 3770K). This is of course well above what AMD specified the chip would do when giving motherboard manufacturers numbers to work to, thus anyone who didn't over engineer their board has a problem.

Again, no it isn't it's AMD's fault, and what does LGA2011 have to do with anything? :S

No,its MSI's fault. They tested the CPUs beforehand,so should have known the weakness of their designs.

Perhaps you should look at the 970-G46 again.

So,again where are the MSI 4 phase VRM designs for Intel 130W TDP CPUs??

Lets look at the MSI website again:

http://www.msi.com/product/mb/#/?sk=Socket 2011

MSI don't seem to use them.

So,basically an FX8350 which in terms of power consumption is more like a a higher end Core i7 six core CPU,is now OK to run on very leanly specced power delivery system.

Funny,that their socket 2011 motherboard designs seem to not agree with that.

However,again internal testing should have identified the problems,before the CPU lists were published.

GB making an error, the point was MSI are not the only manufacturer to have VRM issues.

In this case,it seems very weird you are adamant to blame AMD for the issues with the 970-G46A but not Intel for the problems with the Gigabyte X79 motherboards.

So,if the 970-G46A is an example of AMD lying,then surely the Gigabyte X79 problem is the fault of Intel using your logic??

Addressed previously in the thread, its Chinese whispers due to the customer support guy relaying info from R&D.

So R and D,don't know the difference too or is the chap just incompetent??

Well if AMD hadn't given out incorrect specs for the FX-8350 to MSI then I'm sure MSI would have either designed the board for an FX-3850 or just rated it for an FX-6300...

Or alternatively,MSI has just overrated their motherboards AGAIN.

Moreover,why are they listing the FX8350 for the 970-G46A:

http://www.msi.com/product/mb/970A-G46.html

Surely,if during testing,they should have seen this "issue" and then said the 970-G46A only supports 95W TDP CPUs.

Its listed as an overclocking capable motherboard. However,it throttles with a CPU at stock clockspeeds. Since MSI don't list any disclaimers to what CPUs which can be overclocked,it is more shoddy advertising from MSI,which should have been removed once the "issue" was seen. It makes me think even overclocking a 95W Athlon II X4 might cause the motherboard to have problems.

They didn't and MSI basically told porkies for marketing purposes,so they could sell a £55 motherboard as being "8 core" compatible with no addtional R and D cost.

So at this point I don't agree with you. Any company should have CPU lists based in reality and verified by testing BEFORE being listed on their websites.

This talks volumes about the shoddy QC and QA MSI have,which is now creeping into their graphics cards.

So,we will have to agree to disagree on this one.

 

[ubersonic]

No,because it appears you are adamant this is the fault of AMD.

Okay, lets use an analogy, say you make motherboards and AMD tell you that their new CPU will draw up to 300w at peak (made up lol numbers ofc). You build your entry level motherboards to cope with a CPU drawing up to 300w at peak. However when the boards and CPU are in production it then turns out that the CPU really draws 500w at peak and if people burn test it it will blow mosfets. You then have to use software trickery to make the CPU throttle on your board to avoid that happening.

Who's fault is it?

 

[wolvers]

So basically AMD said the CPU will use up to X power, but in fact it turns out to use X+Y power,

Do you really still not understand the difference between TDP and power usage? We've been through it enough times in this thread already, AMD are not stating that these CPUs are drawing 125w of electrical power.

 

[JonSmith]

It makes me think even overclocking a 95W Athlon II X4 might cause the motherboard to have problems.

Funny you should say that, my old AII x4 635 ran fine at 3.5GHz in my gigabyte 770t-UD3. Exact same settings in an MSI 770 motherboard? CMOS reset at above 3.2GHz.

 

[ubersonic]

Do you really still not understand the difference between TDP and power usage? We've been through it enough times in this thread already, AMD are not stating that these CPUs are drawing 125w of electrical power.

The post you quoted had no mention of TDP, in fact if you read back through the quotes/replies you will see I personally point out that TDP and power usage are not the same:

In case people have become confused by all the cross talk of TDP and power draw in this thread, were not talking about a CPU that's rated at 125w TDP and consumes >125w power, that's a given, we're talking about a CPU that is officially rated and marketed as being 125w TDP yet is actually ≥140w TDP and consumes nearly 200w under full load.