AMD Throwing the Gauntlet at Intel for releasing biased & unreliable benches.

[OneMoar]

I don't see it. CPU TDP has pretty much remained the same throughout my knowledge, through different market segments.

what you aren't factoring is performance per watt and this is where amd is getting its assbeat
amd cpus are garbage from a performance per watt standpoint and thats really the only thing that matters these days
the more efficient your architecture the better the performance they go hand and hand
amd is no longer the underdog they are just a old crippled dog that shits all over the house and should be shot .... I am so tired of the debate amd-cpus= shit period they have nothing to offer in the cpu market

 

[Bill_Bright]

I agree (though there are more 65W CPUs) wattage really does not really apply here.

 

[lilhasselhoffer]

what you aren't factoring is performance per watt and this is where amd is getting its assbeat
amd cpus are garbage from a performance per watt standpoint and thats really the only thing that matters these days
the more efficient your architecture the better the performance they go hand and hand
amd is no longer the underdog they are just a old crippled dog that shits all over the house and should be shot .... I am so tired of the debate amd-cpus= shit period they have nothing to offer in the cpu market

Slow clap.

Thank you, for completely missing the point. Please note that it isn't just you missing the point, but such a brazen opinion deserve a response.


What this is about is a few tests run on two pieces of mobile hardware. Please, look back at the link.


Argue that Intel is better until you're blue in the face. Argue that the tests are unfair, and that's the only reason AMD loses. Neither of those points matter.

Can we sit down, and really ask what this is? AMD is arguing that a test is unfair, but their conclusion is that even their "fair" test results in them being beaten. They've chosen a specific market segment, defined "similar" hardware to test on, and they're not claiming superiority. What AMD has claimed is that they aren't as far behind Intel as Intel's figures claim. They've demonstrated it by taking two separate tests, with separate goals, and said Intel isn't fair. It's a definite step up from claiming that their processors somehow handily beat out Intel.


This is AMD PR. It's BS, and taking it to the level that we have is...bewildering. AMD has admitted they're behind. They've not demonstrated any real data that denies that. AMD is desperately trying to get PR, and they've resorted to lighting themselves on fire to do it. Can we please let their stupidity burn out, rather than ignite another flame war between fanboy bases? Barring that sanity, can we all agree that this isn't AMD claiming superiority, just them claiming to not be completely out of the race by arbitrary standards they set? No matter what side of this debate you're on, can we at least agree that AMD PR no being fired en mass is a miracle?

 

[cdawall]

I don't see it. CPU TDP has pretty much remained the same throughout my knowledge, through different market segments.

P4 670 was the top tier chip in its day. Maximum power usage was 148 watts "115w TDP", The QX9775 150w TDP; a 6700K is a 91w CPU. The 6700T/TE uses 35w. Wattage has come down, also remember under normal usage intel's tend to be well under TDP unlike CPU's of old. Power usage is way down compared to days of old on CPU's.

 

[hat]

what you aren't factoring is performance per watt and this is where amd is getting its assbeat
amd cpus are garbage from a performance per watt standpoint and thats really the only thing that matters these days
the more efficient your architecture the better the performance they go hand and hand
amd is no longer the underdog they are just a old crippled dog that shits all over the house and should be shot .... I am so tired of the debate amd-cpus= shit period they have nothing to offer in the cpu market

I'm not denying that performance per watt hasn't gotten better, merely pointing out that the watts are still roughly the same. On to your other comment, yes, AMD is behind. Yes, their processors are slower, and use more power doing so. What's worth noting, however, that benchmarks and high performance computing don't matter to the vast majority of users. We are enthusiasts. Would I use AMD in any of my systems, if I had the money to do it all over again? Probably. Not in my main machine, given the chance, because now I do a lot of encoding. I rip DVDs, and in doing so employ a number of complex high quality processes which my stock i5 2400 outperforms my overclocked Athlon II x4 by more than double in some cases. No other machine in the house is tasked with such work. AMD would work fine for everything else, on a lower budget. The next most demanding thing is an el gato capturing PS4 gameplay which an overclocked Q6600 struggles with, but even a stock FX8320 would handle with ease (those bastardized psuedo-cores they tacked on aren't useless in everything). To be able to handle that with an Intel, I would need at least an i5 which would cost a good bit more.

 

[Schmuckley]

Crying about Sysmark? What about XTU?
Where's the AMD XTU? yah..
This is AMD's year..or not.
I you see my avatar..yeah..I'm not biased.I go where the performance is.
I have have been using both Intel and AMD AND Cyrix for a long time.

 

[Bill_Bright]

What AMD has claimed is that they aren't as far behind Intel as Intel's figures claim. They've demonstrated it by taking two separate tests, with separate goals, and said Intel isn't fair. It's a definite step up from claiming that their processors somehow handily beat out Intel.

This is AMD PR. It's BS, and taking it to the level that we have is...bewildering.

You correctly point out that AMD is accurately noting their CPUs are still being beaten, but their complaint is that it is just not as severely as Intel want's everyone to believe. Then claim that is BS! It is not BS. AMD is being truthful. Bewildering they publish that, maybe, but it is not BS.

I see this (and I may be dating myself but who cares?) as the old Avis car rental commercials where they readily admitted to being in second place behind Hertz. They were being truthful. That said, I am not claiming AMD is trying harder, as Avis did.

 

[newtekie1]

Sorry, but I don't agree with that at all. They are quieter and more efficient. 10 years ago, even with mild overclocking, the OEM coolers (with AMD or Intel) would not be sufficient. Today, OEM coolers are able to support even moderate overclocking - with properly configured case cooling, of course.

On the left is a 10 year old Intel stock heatsink I received with my Pentium D 805. On the right is a current highest end stock Intel heatsink I received with my 4790K. Please, explain to my how the one on the right is significantly better, more efficient, than the one on the left, and isn't in fact worse.

 

[TheGuruStud]

Intel has been doing this even when AMD left them in the dust. They just pay off the mags and review sites. What else is new?

Only fools believed a pentium 4 was 25% faster than the athlon 64 LOL. Just like a fool will believe that intel wins by 200% in every program ever, today.

Intel should have forfeited their entire cash supply to AMD for rigging the market. They were slapped on the wrists.

 

[cdawall]

Intel has been doing this even when AMD left them in the dust. They just pay off the mags and review sites. What else is new?

Only fools believed a pentium 4 was 25% faster than the athlon 64 LOL. Just like a fool will believe that intel wins by 200% in every program ever, today.

But they are 200% faster in horribly threaded CPU bound games that only utilize a single core.

 

[lilhasselhoffer]

But they are 200% faster in horribly threaded CPU bound games that only utilize a single core.

You mean most of the games produced, ever.

You're going to have to help me here, because it seems like most people are indulging in revisionist history here. For most of the history of video gaming the hardware on the console side has been single core. Developers code for consoles, as that's the "largest" player base. Likewise, computer developers haven't really been on the multiple threading band wagon, because of the huge step up in complexity associated with writing threaded applications. If that isn't apparent, let me list off two big publishers that have released games (and underlying engines) that will not run if too many threads are enabled (and best yet, have been put out in the last decade).
Bethesda - Fallout 3/New Vegas (look up the iNumHWThread .ini tweak for confirmation)
Activision - Prototype (Will not run, period, on a system that has more than 2 threads. Ran fine when only 2 cores were enabled in BIOS)


What you're telling me is that it's unfair to compare single threaded performance. It's unfair because the tests are rigged. Let's be real, the software is what matters to people playing video games. If the software is crippled to run on only one, or at best two threads competently, then why is it unfair to test single threaded performance? Likewise, it's not really reasonable to ask for a test of a complex system as a whole, and interpret relative performance of your component in that system. That's what AMD is asking for with the alternative test, they want to put standardized delays into the mix (read: SATA communication, memory access, etc...) so that their deficiencies are less obvious. Sleazy tactic, especially when your entire data set consists of two pieces of hardware and two tests. 4 data points, and they accuse Intel of unfair practices. Intel deserves to be challenged, but this is more absurdist comedy than viable challenge.

Let's give credit where it's due. AMD took a bold step with Bulldozer, and entirely failed to get enough development behind their unique improvements. They managed to pull a Netburst out of previous success, and have decided to ride that failure all the way to Zen. While I can respect sticking to your guns, AMD is well past the point where it's obvious that they bet on the wrong horse. Let's let this bit of idiocy die. Let's tell AMD that we're smart enough to realize this is PR crap, and they haven't got a leg to stand on. Let's ask them to make sure that Zen rectifies their shortcomings, rather than continues this self destructive pattern of absurdist claims that don't match with reality. Can we please not make this another flame war, based on 4 distinct points of data which even AMD admits prove they have an inferior product?

 

[TheGuruStud]

What? I played prototype (smoothly might I add) on my 4 and 8 core AMDs... Lol

 

[Filip Georgievski]

Hahaha, i played Prototype 1 and 2 on a Core 2 Duo E7500 and my old 6770 with no problem what so ever as well.
That is a laught coming from an AMD CPU.
Try playing Witcher or Dying Light.
I'm not saying AMD CPUs are bad cause for that money, they are worth it, just that Intels equivalents beat them at most aspects.
Dont get me wrong, i had an Athlon 64 (i can get it out and send you pictures for proof) and it beat the shit out of a Pentium 4.
But, now Intel is on the top, and that is that.

PS: Fanboys, pls go back to your gaming and bemchmarks, leave this discusion to the more experienced consumers and people who know what they are talking about.

 

[Tsukiyomi91]

I too used AMD chips before moving to Intel for the sake of performance & efficiency. Price wasn't exactly the issue for me... sure AMD chips are cheap (no puns intended), but the problem they have is heat generation, whether you use stock or aftermarket coolers & they tend to eat a lot of juice in order to keep up with Intel chips which only used around a fraction of what AMD requires. Real world usage however, sure there isn't much of a difference but paying your monthly utility bills is a pain in the ass if your system is all-AMD that runs the whole day. It's also bad if you have all-year long summer heat like over here in MY... So which do you prefer? super-cheap hardware but crazy high monthly fees for your electricity bills OR premium hardware but easy on the wallet/bank electricity bills? I would choose the 2nd option if I were you...

 

[lilhasselhoffer]

Hahaha, i played Prototype 1 and 2 on a Core 2 Duo E7500 and my old 6770 with no problem what so ever as well.
That is a laught coming from an AMD CPU.
Try playing Witcher or Dying Light.
I'm not saying AMD CPUs are bad cause for that money, they are worth it, just that Intels equivalents beat them at most aspects.
Dont get me wrong, i had an Athlon 64 (i can get it out and send you pictures for proof) and it beat the shit out of a Pentium 4.
But, now Intel is on the top, and that is that.

PS: Fanboys, pls go back to your gaming and bemchmarks, leave this discusion to the more experienced consumers and people who know what they are talking about.

?

Maybe you've cited this incorrectly, but the E7500 was a dual core processor. Note above, where I said 2 cores worked.

I can't tell if you didn't read, or I'm missing the point. I tried to play Prototype on a system with a 3930k and a 7970 GPU, both of which should more than handily meet requirements. It would get part of the way through the opening cinematic, then hang. I figured it was some bass-ackwards setting, so off to google I went. After half an hour the only fix I could find was related to core count. Popped into the BIOS, disabled HT and 4 cores, and restarted the system. Everything was slow, but I turned Prototype on, and everything worked. Tell me, given the provided information how exactly am I supposed to come to any conclusion but threaded performance for Prototype was...poor?


If Prototype isn't your particular thing, do read up on Fallout. There are a number of tweaks to the .ini file that are necessary to get it running on a modern system. One of the biggest stability tweaks is to set the amount of threads available to the game to two, and set those particular threads as highest priority. Fallout 3 on windows 7 is an absolute mess otherwise. If my word, and the huge list of people online experiencing that issue isn't enough, then let's ask Valve. It might have gone unnoticed, but Valve has recently added information to Fallout 3 that suggests it doesn't run well on a modern OS. Their exact words are: "Notice: Fallout 3 is not optimized for Windows 7 and later."

My point was simple, threaded games is a relatively new concept. Most games have been produced prior to the effort to thread programs. To suggest it's unfair to test single core programs is insane. It's advantageous to Intel, but that's because Intel didn't chain their products to the need for programmers to fundamentally rework how they code. AMD did. AMD put themselves where they are now, they admit to producing a poorer performing CPU, but they want to change the testing conditions to mitigate that difference. It's crap, and anyone with an ounce of reason should be calling them out. 4 data points with two pieces of hardware isn't justification for Intel cheating, it's justification for AMD to generate PR on the lead-up to their product release.

Unfortunately, Zen is about 10 months off. AMD is generating this fanboy s***storm for no reason other than to stay relevant. They want to sell off their remaining stock of under performing (though budget friendly) CPUs before Zen rolls out and obliterates the ability to move them. Somebody at AMD did the math, and discovered you'd need 10 months to sell existing stocks. Welcome to reality, where the cynical view is generally correct, but AMD banks on the fact that zealots will start flame wars instead of trying to discover their thinly guised attempt to shift inventory before it's completely irrelevant.

 

[cdawall]

I normally early adopt amd stuff, but Zen will wait until reviews.

 

[trog100]

once amd tweaked the sleeping tigers tail.. but the tiger woke up and with one swipe of its paw put the upstart back in its place.. nothing has changed since then..

how much longer amd can survive is anybodies guess.. maybe not that much longer even though the market needs them..

if they can offer the same (mid range) performance for less money they may stand a chance but if they do its purely down to intel helping them.. intel could if they so wished lower their prices at any time.. its a rigged market and intel set the rules.. in a way intel needs amd which is probably the only reason amd has survived this long..

when you have only two players players true competition dosnt exist.. the leading player sets the rules end of story..

trog

 

[Bill_Bright]

@newtekie1 - sorry, but I don't see any images.

That said, not sure pointing out one pair of images makes the case for either of us. There are many variables that come into play here. When it comes to heat sink materials, one tiny change in the composition of the metal alloy can significantly change the efficiency of the heat conduction. One tiny change in the fin shapes, thickness, size and number can make a significant change in the efficiency. Raw materials are much more pure. Manufacturing techniques have greatly improved in recent years to ensure the mating surfaces are free of imperfections and are more perfectly flat. And OEM TIM has much improved too. All these factors together improve cooling efficiency.

In the olden days, we use to "lap" heat sinks and CPU dies to ensure perfect flatness. That is no longer needed.

Fan technologies have greatly improved too. R&D has gone into the aerodynamic shape of the blades (they are tiny wings, after all) to ensure they grab and push more air, instead of just chopping at the air. Better bearings are used too. Better design blades and precision bearings decrease noise too.

And CPUs are much more efficient too. Plus, cases today don't come with a single 80mm case fan as cases 10 years ago commonly did - keeping in mind it is the case's responsibility to provide a sufficient supply of cool air running through the case.

So looking at a couple pictures proves nothing.

IF the OEM coolers were as bad as you want us to believe, there would be 100s of millions of overheating computers out there. And that is just not happening.

Again, if doing extreme overclocking, if your case cooling is inadequate, of if you are just seeking bragging rights (or you want silent running for a HTPC), then by all means go for an aftermarket cooling. But if you are just using standard clocking, I say try the OEM cooler. To be sure, I am NOT saying OEM coolers are the panacea for CPU cooling. I am just saying OEM coolers are not crap.

Another thing I've said before - keeping our CPUs adequately cooled is, no doubt, absolutely essential for stable operation and long life. But cooler does not automatically mean better. A CPU running at 35°C will not be more stable, perform better, or last longer than a CPU running 45°C or even 55°C (or even higher, for some CPUs).

You really need to ask yourself, "why would Intel or AMD provide coolers (and warranty them too) that failed to keep their CPUs adequately cooled?" How could they and not go bankrupt?

I've been doing IT tech support for a living for over 40 years and the idea that OEM coolers are inadequate is just not supported by the facts! It is myths perpetrated by aftermarket cooler makers, enthusiasts who believe everyone should follow their lead, and their blind followers.

So I say again, give the OEM coolers a chance. Make sure you have adequate case cooling. Keep your case clean of heat trapping dust and you just might be surprised at how efficient, and quiet the OEM coolers can be.

If still not satisfied, then go for an aftermarket cooer.

 

[cdawall]

IF the OEM coolers were as bad as you want us to believe, there would be 100s of millions of overheating computers out there. And that is just not happening.

They are overheating. They just throttle back when it happens. In fact they perform so badly most OEM's use aftermarket heatsinks even on the i3's

 

[Bill_Bright]

That's just BS. So you are claiming now that 100s of millions of user PCs are not running at full speed because they are overheating and the users are not aware of this so they are keeping quiet?

Yeah right. Nice try.

 

[cdawall]

That's just BS. So you are claiming now that 100s of millions of user PCs are not running at full speed because they are overheating and the users are not aware of this so they are keeping quiet?

Yeah right. Nice try.

I wouldn't know I just work in a tech shop and deal with it on a day to day basis. They can run normal clock speed fine it is the turbo mode that most of them down clock from. I would also love for you to get me a "normal user" and have him tell me the difference between a CPU that is throttling down to 2.4-2.6ghz vs 3.9

and again as I already said I rarely see an intel OEM HSF in a prebuilt. Most of them have an aftermarket cooler that is better than what intel provided.

 

[cdawall]

Let's add pictures to prove the point.

This is an unoverclocked dell with an i7 6700k. Mind telling me which heatsink is better this or the one @newtekie1 posted? And this is fucking dell, one of the most returned, worst built consumer level products on the market.

 

[Bill_Bright]

Oh yeah. That totally proves everything. 100s of millions of users out there are all a bunch of dumb bunnies.

I wouldn't know I just work in a tech shop and deal with it on a day to day basis.

And I own a custom PC and consulting business. I have government and business contracts to support their computers and networks. And I know better than to assume what I see coming into the shop represents what is out there in the real world.

It is clear you are convinced and so are unwilling to even to see for yourselves.

Oh, and BTW, most (if not all) CPUs today toggle down in speed because they don't need to run at full throttle. Not because they are running too hot. CPUz will show anyone that.

If you want to use an aftermarket cooler - go for it. But when giving advice, if the user will not be doing extreme overclocking, it is bad advice to automatically tell them to spend extra money on an aftermarket cooler if their CPU comes with an OEM cooler.

And for those of you who automatically recommend side firing coolers like the CM 212, think twice. Motherboard designers intentionally surround the CPU socket with other heat generating and heat sensitive devices so they too can take advantage of the outward spreading air flow created by the downward firing OEM coolers. Aftermarket side firing fans do not provide such needed cooling as the fan is up high and blows in only one direction.

Don't take my word for it. See the * note here where it says (my bold underline added),

* For cooling the CPU and its surrounding components, please install a CPU cooler with a top-down blowing design.

This is just one example from ASRock. Other motherboard makers have posted similar warnings. So, if your CPU does not come with an OEM cooler, or you just choose to use an aftermarket cooler, I recommend you get one that fires in the same orientation as the OEM coolers the designers anticipated to ensure your sensitive motherboard components receive the cooling they need too. This surely will help provide better stability, regulation and component longevity on motherboards that rely on that flow of air.

Now I see no point in discussing this further - unless you can provide real evidence supporting evidence that 100s of millions of CPUs out there have throttled down because the OEM coolers fail to keep them cool.

 

[cdawall]

Oh yeah. That totally proves everything. 100s of millions of users out there are all a bunch of dumb bunnies.


And I own a custom PC and consulting business. I have government and business contracts to support their computers and networks. And I know better than to assume what I see coming into the shop represents what is out there in the real world.

I work in the real world... Next time you have a business computer do me a favor and pop the side panel off. Guarantee it isn't a stock intel HSF, now tower cooler vs top down that's purely on manufacturer. I couldn't care less about custom built machines that is not even close to a majority market share. Want to see which designs work? Pop open an optiplex, thinkcentre or prodesk. They do not have OEM intel HSF or OEM AMD HSF's in them for a reason. The stock coolers do not have the ability to cool a machine at max turbo settings and all of this is a known well documented thing.

Oh, and BTW, most (if not all) CPUs today toggle down in speed because they don't need to run at full throttle. Not because they are running too hot. CPUz will show anyone that.

I know how throttle stop and CnQ works...I don't need you to explain them.

Now I see no point in discussing this further - unless you can provide real evidence supporting evidence that 100s of millions of CPUs out there have throttled down because the OEM coolers fail to keep them cool.

This doesn't apply to the vast majority machines because they don't use the crap intel cooler...I think I mentioned not even the OEM's use them what 3 times now?

Also why did you bring up anything about tower coolers? Literally just said it was a better cooler than the intel one. Don't really need another one of your random off topic rants trying to change the subject...

 

[trog100]

throttling down on turbo settings is back to front logic.. the base clock is the guarantee clock turbo is an extra.. if conditions are right you get the turbo if they aint you dont get the turbo or at least not so much of it..

and for f-cks sake it aint in intels interest to ship stock coolers that will cause customer problems.. dont get me wrong here.. they will throttle down temp wise if they hit 100 C which many will running daft things like prime95 or even intels own burn in test..

but normal people dont run daft things like prime95.. only enthusiasts do that.. and intel make adequate provisions for that..

view turbo boost for what it is.. a little extra over and above if conditions (the software being run being one of them) are right and it make more sense..

a graphics card behaves in the same way.. you only get the max claimed boost if conditions are right.. sometime they are.. sometimes they aint..

trog