Dont pass judgement on bulldozer yet...

[frozentundra123456]

Power consumption of a 2600K suicide OC run != That of 2600K @ 24/7 safe overclocks.

Many reviewers can't even push their BD past 4.6GHz. What's with this new wave of AMD defenders lately?

I dont get it either, why so many people try to defend bulldozer. I admit it is possible that in a year or so, it will be improved with better IPC and lower power usage. But Intel will not be standing still over that time either.

So at best, it might become competitive if every attempted improvement works out optimally and if they reduce the price also. But you are still stuck with a 2B transistor chip that will have to sell at a discount.

And granted, some software may become more optimized for Bulldozer, but I would think that you would want to be efficient on the software that is prevalent when the chip comes out, not something that may or may not be used in the future.

 

[Ferzerp]

OMG if I see another new reg defending Bulldozer I'm going to scream.

Can't we get a subforum, or a consolidated thread or *something*? So that we can have the old CPUs and Overclocking back.

 

[Idontcare]

Power usage? Thats bull...

I've read a review at rag3d(tweak town) if im not mistaken were they put a 5.2ghz 2600k with 3 hd6970 and FX-8150 4.76ghz with 3 hd6970 also, and power usage is 703W for 2600k and 723W for Bulldozer.

Link :http://www.tweaktown.com/articles/4...crossfirex_hd_6970_x3_head_to_head/index.html

It's not that different at all, take in consideration it's freaking 8-cores not 4+ HT.

Those 2600K power numbers look like BS to me.

http://forums.anandtech.com/showthread.php?t=2195927

This is total power consumption (system and CPU) at full load with IBT all the way up to 5.0GHz:

At 5GHz, loaded with LinX for maximum power-consumption, my 2600K consumes 225W, my system consumes 125W (including a GTX460 1GB OC'ed to 905MHz @ 1.1V w/bios mod).

Going from 5GHz to 5.2GHz is not going to increase the CPU power consumption by another 400W. It's just not. Something is borked with the tweaktown results, especially when you consider they aren't even loading it with IBT, that's just 3Dmark11 there.

 

[Maximilian]

5-stages of grief:

1. Denial — "AMD is fine, they've been down before, they'll bounce back!"
Denial is usually only a temporary defense for the individual.
2. Anger — "WTF is this BD BS? It's not fair!"; "How can this happen?"; '"Who or what is to blame? Is it the compilers? Those lazy good for nothing programmers? JFAMD?"
Once in the second stage, the individual recognizes that denial cannot continue. Because of anger, the person is very difficult to care for due to misplaced feelings of rage and envy.
3. Bargaining — "I'll buy them anyways, gotta support competition."; "Where's the undervoltage testing? Surely there is a niche for this product where it WINRAR's the day..."; "Just wait till Piledripper or SteamingPile or Constipator are released, AMD has promised it will do better next year!"
The third stage involves the hope that the individual can somehow postpone or delay the inevitable. Psychologically, the individual is saying, "I understand this sucks, but if I could just do something to convince myself to just wait longer..."
4. Depression — "I'm so sad, why bother with AMD anymore?"; "I'm just going to buy an Intel SB soon anyways so what's the point... What's the point?";
During the fourth stage, the ardent AMD supporter begins to understand the certainty of AMD's situation. Because of this, the individual may become silent, refuse to login to their favorite forums and spend much of the time crying and grieving. This process allows the individual to disconnect from things they once held near and dear. It is not recommended to attempt to cheer up an individual who is in this stage. It is an important time for grieving that must be processed.
5. Acceptance — "It's going to be okay."; "AMD can't fight it, they may as well prepare for it."; "NV will probably buy them anyways..."
In this last stage, individuals begin to come to terms with the inevitable, they move on with life, pleasantly surprised to find that their newly acquired 2600K does not give their kids leukemia despite what they read on AMDZone...

​

From IDC's post in the other thread it looks like the OP is at stage 3. Maybe when windows 7 is optimized, maybe its a firmware/driver problem yada yada yada.

Dont worry OP, remember the 2500k is still a great CPU, that thought will help you through stage 4

 

[SithSolo1]

Power usage? Thats bull...

I've read a review at rag3d(tweak town) if im not mistaken were they put a 5.2ghz 2600k with 3 hd6970 and FX-8150 4.76ghz with 3 hd6970 also, and power usage is 703W for 2600k and 723W for Bulldozer.

Link :http://www.tweaktown.com/articles/4...crossfirex_hd_6970_x3_head_to_head/index.html


It's not that different at all, take in consideration it's freaking 8-cores not 4+ HT.

Yes they have similar power draws there. Assuming both are complete stable, here's my issue:

A stock 2600k is 200 Mhz slower than a stock 8150 and still trashes it in games(for me, thats all I care about). That OC'd 2600k is 450 Mhz faster than the OC'd 8150. At the resolutions they would use that system for the 2600k is gonna slap that 8150 down like a red-headed step child. My guess is that you wouldn't need more than 4.2-4.4 on the 2600k to be on par with the 8150 @ 4.76. Dropping that 2600k roughly 800+ Mhz is gonna drop the cpu load power draw substantially.

Going from 5GHz to 5.2GHz is not going to increase the CPU power consumption by another 400W. It's just not. Something is borked with the tweaktown results, especially when you consider they aren't even loading it with IBT, that's just 3Dmark11 there.

Maybe that is total system draw, not just CPU.

 

[Hypertag]

Scream then, people come so hard on BD because of fanboyism.

Yeah, everyone that hates bulldozer is an Intel employee, puppet, or investor ...

 

[Tikerz]

OMG if I see another new reg defending Bulldozer I'm going to scream.

They're probably longtime members of AT, just afraid to use their real account to defend BD.

 

[Hypertag]

Those 2600K power numbers look like BS to me.

Something is borked with the tweaktown results, especially when you consider they aren't even loading it with IBT, that's just 3Dmark11 there.

Good eye. That is how they produced those inaccurate results. The proper method of testing max power draw is furmark for GPUs and prime95 for CPUs.

There is a significant chance that both the bulldozer and 2600k would not pass a prime95 stress test at those overclock levels. That is why proper stress testing is critical, besides getting accurate power usage figures.

 

[paperfist]

I'd like to know why it wasn't tested using DDR3 1866 on most sites. In the benchmarks I've seen with that memory BD is on the level of the i5-2500k.

 

[superccs]

Wiat what happened to 140W or 125W power envelopes? My mainboard is only good for a 125W CPU, how can a BD be drawing 400W?

 

[BlueBlazer]

OMG if I see another new reg defending Bulldozer I'm going to scream.

Its becoming a new trend.

 

[Nemesis 1]

The people who are defending Bulldozer now are probably the same people who bashed Prescott back in the days, assuming they were around back then. It's funny because Bulldozer and Prescott share alot of traits, like massive power consumption/heat and decent performance in encoding/rendering type apps while sucking in games, compared with their competition.

The thing is Prescott came after the P4C . So Intels screwed the pooch even harder . This is AMDS first try . If AMD gets its front end properly configured it has promise. If I ran AMD I would have run away from Intels AVX it was a trap . AMD jumped right into it . As intel showed AMD only part of the AVX plan. AMD should have stayed with its orginal plan instead of switching to intels. But they did know it befor releaseing BD . Now that BD has released it they are stuck with it . Developers are going to develop for intels AVX. or lose . Intels market preseance assures that . If a company is tring to make money . Haswell will open all eyes.

 

[carnage10]

So pretty much the OP is asking us to be patient, give AMD the benefit of the doubt and have faith that they will deliver in a timely manner, while AMD remains completely tight lipped and gives no guarantees or indicators otherwise?

De je vu anyone?

We've been hearing the same thing from AMD hopefuls for years now, while they ignored the telltale signs of a flop in the making, and optimistically denied themselves excellent all round performance and thermals TODAY (well for 10 months actually) by choosing not to purchase a SB system and wait and see what BD would bring.


There is no magic bandaid fix for BD. Maybe if they get Piledriver out in a timely manner (which is asking alot) and can get the thermals down and performance up, maybe by that time software will start to be optimized for the arch, making it more competitive with Intels offerings in the future. But BD in it's current iteration doesn't have enough going for it to make it a good choice over SB TODAY.


As much as id love to see AMD release the next Intel killing Athlon 64, their CPU division just doesn't have track record or credibility to meet deadlines or meet reasonable expectations. I fear Intel will never give up the performance crown

 

[ocre]

Yes they have similar power draws there. Assuming both are complete stable, here's my issue:

A stock 2600k is 200 Mhz slower than a stock 8150 and still trashes it in games(for me, thats all I care about). That OC'd 2600k is 450 Mhz faster than the OC'd 8150. At the resolutions they would use that system for the 2600k is gonna slap that 8150 down like a red-headed step child. My guess is that you wouldn't need more than 4.2-4.4 on the 2600k to be on par with the 8150 @ 4.76. Dropping that 2600k roughly 800+ Mhz is gonna drop the cpu load power draw substantially.



Maybe that is total system draw, not just CPU.

a typical misinformation attempt to misdirect!

It is mostly the GPUs power thats making the difference with the benchmark they used. And since the 2600k is not holding back the CF as much as the BD, its only natural the 2600k system is using more power. Its GPUs are pushing harder and therefor the total power draw is higher.

There are plenty of CF and SLI reviews out there that show clearly BD isnt good for scaling, it holds back multi gpu performance. The intel CPUs walk all over BD is multigpu setups. Here is one with the i7 920 and look at how bad the BD does in CF:

http://alienbabeltech.com/main/twen...-6970-6970-crossfire-vs-phenom-ii-and-core-i7


with the 2xxxk it would be even worse.

 

[jhansman]

Too late dude. The Intelites have been dancing on AMD's grave for months now. Classy bunch, they....

 

[ErickMaga]

Going from 5GHz to 5.2GHz is not going to increase the CPU power consumption by another 400W. It's just not. Something is borked with the tweaktown results, especially when you consider they aren't even loading it with IBT, that's just 3Dmark11(forgotten about the GPU's?)(3) there.

Remember there are 3x HD6970, 3Dmark is used to similute games stiuation, in this case, games power draw.

Both are alike. Thats all i care, i use 3xHD6950@70, so when people saw BD consumes so much power, power bill will quadruple blablabla, consumes two times what SB consumes..., i say its bull..., my TX950 will handle my setup just fine.

I'm not defending BD is better than SB, just saying power draw is about the same, i've seen many people saying that bulldozer is bad because of its immense power draw.

 

[Idontcare]

Remember there are 3x HD6970, 3Dmark is used to similute games stiuation, in this case, games power draw.

Both are alike. Thats all i care, i use 3xHD6950@70, so when people saw BD consumes so much power, power bill will quadruple blablabla, consumes two times what SB consumes..., i say its bull..., my TX950 will handle my setup just fine.

I'm not defending BD is better than SB, just saying power draw is about the same, i've seen many people saying that bulldozer is bad because of its immense power draw.

Then power consumption for the system really should be normalized in performance/watt and not shown in absolutes.

Do both the 5.2GHz 2600K and 4.7GHz BD systems get the same 3DMark11 scores with that GPU setup?

Having comparable power-consumption but not having comparable performance is relevant to the discussion.

 

[Rvenger]

Power usage? Thats bull...

I've read a review at rag3d(tweak town) if im not mistaken were they put a 5.2ghz 2600k with 3 hd6970 and FX-8150 4.76ghz with 3 hd6970 also, and power usage is 703W for 2600k and 723W for Bulldozer.

Link :http://www.tweaktown.com/articles/4...crossfirex_hd_6970_x3_head_to_head/index.html

It's not that different at all, take in consideration it's freaking 8-cores not 4+ HT.

All this proves to me is that the 2600k can o/c 500mhz faster and still consume less power than a Bulldozer. I really don't know what this proves about BD.

 

[ErickMaga]

Then power consumption for the system really should be normalized in performance/watt and not shown in absolutes.

Do both the 5.2GHz 2600K and 4.7GHz BD systems get the same 3DMark11 scores with that GPU setup?

Having comparable power-consumption but not having comparable performance is relevant to the discussion.

Performance per watt is good to demonstrate efficiency, they definitively dont get the same scores.

But its not the monster people say it is.

Underperformed? Yes.

Too much hype? Yes.

Looses to Phenom II sometimes? Yes.

Lower IPC ? Yes.

Huge power hog? I think not, there are 8 physical cores, if you look to the power consumption of i7 920 OC'ed, its about the same, but with twice as much cores.

Even the Gulftown with 32mn consumes about the same with 2 less cores.

I'll buy it because i believe it will perform better than my Phenom II @4.1ghz, and OC more than it does.

I'll even disable 1 core in each module and see what can i do with that, if its better than phenom II x4 and if it OCs more.

 

[Maximilian]

Performance per watt is good to demonstrate efficiency, they definitively dont get the same scores.

But its not the monster people say it is.

Underperformed? Yes.

Too much hype? Yes.

Looses to Phenom II sometimes? Yes.

Lower IPC ? Yes.

Huge power hog? I think not, there are 8 physical cores, if you look to the power consumption of i7 920 OC'ed, its about the same, but with twice as much cores.

Even the Gulftown with 32mn consumes about the same with 2 less cores.

I'll buy it because i believe it will perform better than my Phenom II @4.1ghz, and OC more than it does.

I'll even disable 1 core in each module and see what can i do with that, if its better than phenom II x4 and if it OCs more.

So out of the 5 things you mention it loses in 4 of them.

Bulldozer is absolutely terrible from any perspective, it cant be directly compared to gulftown as that CPU has 6 full cores, bulldozer has 8 smaller weaker cores.

Plus saying that bulldozer a brand new architecture on a new smaller process has comparable OC'ed power consumption to the power hog i7 920 from 3 years ago on an older 45nm process is somehow a good thing? How do you figure that? Also same thing with the core comparison applies here, apples to oranges, 1 bulldozer core != 1 nehalem core.

 

[sandorski]

At this point in time, I think we can all agree that BD doesn't seem attractive to most AT'ers/Gamers. Win8, Software updates, improved Production Process, and/or new generations of BD may change that though. Time will tell.

I'm an AMD fanboy, but rehashing the current situation ad nauseum is just annoying. Especially in New Thread after New Thread. If there's some kind of developement, like a Software/Windows Patch, BIOS update, or new Stepping that improves the situation, that would seem to warrant a New Thread.

 

[Idontcare]

Performance per watt is good to demonstrate efficiency, they definitively dont get the same scores.

But its not the monster people say it is.

Underperformed? Yes.

Too much hype? Yes.

Looses to Phenom II sometimes? Yes.

Lower IPC ? Yes.

Huge power hog? I think not, there are 8 physical cores, if you look to the power consumption of i7 920 OC'ed, its about the same, but with twice as much cores.

Even the Gulftown with 32mn consumes about the same with 2 less cores.

I'll buy it because i believe it will perform better than my Phenom II @4.1ghz, and OC more than it does.

I'll even disable 1 core in each module and see what can i do with that, if its better than phenom II x4 and if it OCs more.

You seem to be oddly fixated on "more cores" as the metric of relevance here, promoting it above the ranking of all else.

Has 2 more cores than a gulftown? How is this even remotely relevant when deciding which cpu to purchase? Unless those extra 2 cores makes all the difference in performance.

But this isn't about cores/$, is it? Otherwise you'd be buying silly cheap atoms and bobcats.

Absolute performance, Performance/$, performance/watt, etc are things that are supposed to matter.

"Moar Coars" is a meme for a reason.

 

[frostedflakes]

Power usage? Thats bull...

I've read a review at rag3d(tweak town) if im not mistaken were they put a 5.2ghz 2600k with 3 hd6970 and FX-8150 4.76ghz with 3 hd6970 also, and power usage is 703W for 2600k and 723W for Bulldozer.

Link :http://www.tweaktown.com/articles/4...crossfirex_hd_6970_x3_head_to_head/index.html

It's not that different at all, take in consideration it's freaking 8-cores not 4+ HT.

Well first of all, they probably needed a ton of voltage to hit 5.2 on a 2600K. Reduce the OC 400-500MHz and you could probably use a lot less voltage, reduce power consumption, and still have a CPU that outperforms an 8150 at 4.76. Also, that's power consumption while running 3DMark11. How about power consumption while running a CPU stress test, such as LinX? Would probably tell a completely different story. There's not some conspiracy against AMD here, every review is showing higher power consumption because BD consumes a lot more power.

Anyway, myself and many others aren't necessarily passing judgement on the Bulldozer architecture in general. I do think there's a lot of room for improvement (both on the hardware and software side of things) and in 6-12 months it could maybe actually be competitive. But in its current form (Zambezi), BD is pretty underwhelming. And most enthusiasts aren't going to buy one based on promises of performance improvements in a year or whatever, they're going to buy a 2500K which works really well right now and doesn't need software to be rewritten to take full advantage of it.

 

[ErickMaga]

Bah, i'm done talking to Intel fan boys, sure lets hope intel takes amd out of the x86 market and you can laugh all day long and we all can buy overpriced cpus.

I try to defend BD says its not SO BAD like people say it is, because it really isn't, people read 1 review and take it for granted, i've seen about a dozen review and they all are a little bit different, SB wins in FP calculations but BD win integer.

I game at 3240x1920 and at that res SB or BD will perform just a like.

Bye now i'll play some BFBC2 and kill some intel fanboys/ amd haters.

 

[qliveur]

I refuse to pat AMD on the back for lying to me.

If that makes me an "Intel fanboy", so be it.