HD 7970: Bulldozer vs. Sandy Bridge vs. Nehalem

[Hamlet]

Hey bear!
Computerbase had just the test you need: http://www.computerbase.de/artikel/grafikkarten/2011/test-amd-radeon-hd-7970/18/#abschnitt_eyefinity

 

[cdawall]

W1z is there any chance we could see a crossfire scaling benchmark between the 3 processors you chose? Maybe toss in a Thuban for shits and giggles?

 

[virtue]

meh, would have liked to see the intel CPUs clocked at the same frequency (let's say 4ghz) to show the FPS difference
everyone knows 2500k is superior to the old gen already, comparing it with older gen at higher frequencies was just pointless. basically skimmed the entire article because of that

 

[cdawall]

meh, would have liked to see the intel CPUs clocked at the same frequency (let's say 4ghz) to show the FPS difference
everyone knows 2500k is superior to the old gen already, comparing it with older gen at higher frequencies was just pointless. basically skimmed the entire article because of that

He compared stock clocks vs stock clocks. Its a bit unfair to compare them all at 4ghz if you are going to do that go for normal OC's on them all 3.6-3.8ghz on the 920, 4-4.4ghz on the i5 and 4.4-4.8ghz on the AMD.

 

[JustaTinkerer]

He compared stock clocks vs stock clocks. Its a bit unfair to compare them all at 4ghz if you are going to do that go for normal OC's on them all 3.6-3.8ghz on the 920, 4-4.4ghz on the i5 and 4.4-4.8ghz on the AMD.

very well said.

 

[Dent1]

The Bulldozer's gaming performances is on par with the i7 Nehalem. That is pretty respectable.

 

[FreedomEclipse]

The Bulldozer's gaming performances is on par with the i7 Nehalem. That is pretty respectable.

depends how you see it.... BD's architecture has been in development for around 4years or longer. Nehalem was released back in November 2008 but they had an early preview in IDF sometime in 2007.

Its taken them that long to make a processor that competes with something that was probably finished & ready to ship back in 2007/08?? Hardly the game changer that everyone predicted in the early days.

I think Thuban fills that performance gap pretty well.

 

[Makaveli]

depends how you see it.... BD's architecture has been in development for around 4years or longer. Nehalem was released back in November 2008 but they had an early preview in IDF sometime in 2007.

Its taken them that long to make a processor that competes with something that was probably finished & ready to ship back in 2007/08?? Hardly the game changer that everyone predicted in the early days.

I think Thuban fills that performance gap pretty well.

I agree BD numbers are not respectable at all.

It only closes those gaps when its GPU limited.

And the 920 is a 2008 processor its down right embarrassing if you ask me.

 

[Makaveli]

Thank you for doing this review. It's nice to have reaffirmed with data what I suspected (and hoped really) that I had no need to replace my i7 920. I knew, but looking at actual numbers was reassuring.

Agreed!

I'm currently at 3.8 on my 920 and i'm skipping all of SB and may even skip IB and go straight to Haswell.

 

[Dent1]

depends how you see it.... BD's architecture has been in development for around 4years or longer. Nehalem was released back in November 2008 but they had an early preview in IDF sometime in 2007.

Its taken them that long to make a processor that competes with something that was probably finished & ready to ship back in 2007/08?? Hardly the game changer that everyone predicted in the early days.

I think Thuban fills that performance gap pretty well.

The way i see it is:

I'm not too fused about 2007/2008, only right now. And if I can buy a CPU with gaming performance on par with a i7 Nehalem for cheaper than it I'm all for it.

Ok the Bulldozer had delays, but looking back at this thread people are calling it "junk" and whatnot. To say the Bulldozers gaming performance is weak is to say the i7 Nehalem gaming performance is weak too, which is rediculous.

It's OK to be disappointed, we are all disappointed, but it doesnt mean that disappointment is automatically translated into a bad overal product. It has good (not fantastic) single threaded gaming performance, and fantastic multi threaded performance and on a platform with an upgrade path, this is a good thing.

 

[Makaveli]

meh, would have liked to see the intel CPUs clocked at the same frequency (let's say 4ghz) to show the FPS difference
everyone knows 2500k is superior to the old gen already, comparing it with older gen at higher frequencies was just pointless. basically skimmed the entire article because of that

He compared stock clocks vs stock clocks. Its a bit unfair to compare them all at 4ghz if you are going to do that go for normal OC's on them all 3.6-3.8ghz on the 920, 4-4.4ghz on the i5 and 4.4-4.8ghz on the AMD.

Here you go

http://www.bit-tech.net/hardware/cpus/2011/10/12/amd-fx-8150-review/6

The 920 at 4ghz is still beating the BD chip. The only time it wins is with cinebench and some rendering so it doesn't really change much.

 

[Steevo]

W1zz.....



Bench ALL the options.......lol




You should be able to ascertain the same poor performance from BD at anything but GPU limitied resolutions and stacking up filter options to get the same GPU limited results at lower resolutions won't make the BD any better of a processor.

 

[Makaveli]

The way i see it is:

It has good (not fantastic) single threaded gaming performance, and fantastic multi threaded performance and on a platform with an upgrade path, this is a good thing.

How is having single threaded performance that is lower than the previous chip it is suppose to replace "good"?

 

[TheGuruStud]

Here you go

http://www.bit-tech.net/hardware/cpus/2011/10/12/amd-fx-8150-review/6

The 920 at 4ghz is still beating the BD chip. The only time it wins is with cinibench and some rendering so it doesn't really change much.

I see, so when it's doing work it's designed to do with new instructions disabled (b/c of ICC) and super duper intel optimized benchmark (cinebench) and it still beats intel, then it doesn't matter.

I think you and Tom over at Tom's Hardware would get along nicely. Maybe you can both play in traffic.

I know, AMD can never use SSE3/4, AVX or FMA b/c that just wouldn't be fair.

 

[Steevo]

I see, so when it's doing work it's designed to do with new instructions disabled (b/c of ICC) and super duper intel optimized benchmark (cinebench) and it still beats intel, then it doesn't matter.

I think you and Tom over at Tom's Hardware would get along nicely. Maybe you can both play in traffic.

I know, AMD can never use SSE3/4, AVX or FMA b/c that just wouldn't be fair.

If no real life software uses it then does it matter?


This is like saying a piano goes 0-60 as fast as a sports car, when you drop it from a helicopter.

A great fact, but worthless in real life.

 

[TheGuruStud]

If no real life software uses it then does it matter?


This is like saying a piano goes 0-60 as fast as a sports car, when you drop it from a helicopter.

A great fact, but worthless in real life.

Hmmm, real world apps on an OS that doesn't suck use it, but then again, that OS isn't one of the evil monopolies, so I guess that doesn't count either (linux).

Did you ever stop to wonder why this is? Is it perhaps that intel has been paying off everyone for the last 20 years? During the K7 generation AMD had about 27% market share. It was a very strong CPU, but the athlon 64 was even better and what happened? They LOST market share all the way down to about 15%. Intel doesn't play games. They lie, cheat, bribe, steal, threaten and strong arm anyone in there way. That lawsuit pay out was a joke and so are the results. Almost nothing has changed.

It's true, AMD doesn't help themselves by their position with software devs, but intel actively makes their compiler cripple AMD chips and then pays devs where appropriate to make their CPU look massively better, regardless of its actual perfromance (tons of synthetic benchmarks that are used my the masses). Anywhere you see their logo plastered has effectively been bought.

Oh, and for all you apologists, since when is SSE3/4 and AVX not used in real world apps? FMA is the only new kid on the block (and intel lied about using FMA4, then switched to 3 to screw AMD).

 

[Makaveli]

If no real life software uses it then does it matter?


This is like saying a piano goes 0-60 as fast as a sports car, when you drop it from a helicopter.

A great fact, but worthless in real life.

thank you for pointing out how stupid that point was.

I lol at his personal attack too very childish.

 

[Makaveli]

Hmmm, real world apps on an OS that doesn't suck use it, but then again, that OS isn't one of the evil monopolies, so I guess that doesn't count either (linux).

Did you ever stop to wonder why this is? Is it perhaps that intel has been paying off everyone for the last 20 years? During the K7 generation AMD had about 27% market share. It was a very strong CPU, but the athlon 64 was even better and what happened? They LOST market share all the way down to about 15%. Intel doesn't play games. They lie, cheat, bribe, steal, threaten and strong arm anyone in there way. That lawsuit pay out was a joke and so are the results. Almost nothing has changed.

It's true, AMD doesn't help themselves by their position with software devs, but intel actively makes their compiler cripple AMD chips and then pays devs where appropriate to make their CPU look massively better, regardless of its actual perfromance (tons of synthetic benchmarks that are used my the masses). Anywhere you see their logo plastered has effectively been bought.

I don't know what bridge you been living under but life isn't fair.

 

[TheGuruStud]

I don't know what bridge you been living under but life isn't fair.

What's your address? I want some of your belongings. I'm going to stroll in, take what I want, then leave. Don't bother calling the cops, I will just pay them off.

I love your attitude. Keep supporting crooks. We see how well that's working for the world.

 

[Makaveli]

I didn't create the situation we are in this is way of the world. And it will continue to be this way well after we are both gone.

I however work within the system we have instead of crying over something I cannot control.

 

[cadaveca]

What's your address? I want some of your belongings. I'm going to stroll in, take what I want, then leave. Don't bother calling the cops, I will just pay them off.

I love your attitude. Keep supporting crooks. We see how well that's working for the world.

Intel OWNS x86.

AMD OWNS x64.

They both pay each other for every processor sold. Every Intel CPU sold gets AMD cash, and vice versa.

While Intel and AMD may seem to compete for our wallets, really they are not in direct competition. AMD has a limited number of chips that they can produce, as does Intel. Nearly every single chip AMD produces is sold, and likewise for Intel. AMD originally simply bought a segment of Intel's market, via liscencing, that Intel could not produce chips for.

AMD could have a chip that was 10x as fast as SB, and everyone would be just as pissed as they are at how "slow" bd is, because they'd be so rare that they'd sell for thousands each. While you may not liek the business practices, AMD and INtel are actually PARTNERS in producing computing solutions for the masses. They share so much tech the only real differences between them are their core design, and the silicon it's built on.

Anyway, nice to see the older Intel chips can still ahng out with the new big boys. No need for X58 users to upgrade, if gaming, that I can see. Sure, SB is faster, but X58 is no slouch, either.

 

[ensabrenoir]

^ + 1
The only major difference I found is that is easier to poke fun at amd .....it gets a bigger raise out of their fan base...intellers just laugh with the jokes. But seriously amd fails badly at being intel. Let amd be amd...their execellent at that and let intel b intel.

 

[Dent1]

How is having single threaded performance that is lower than the previous chip it is suppose to replace "good"?

Just because the Bulldozer isnt consistantly faster than it's replacement in singlethreaded activities doesnt mean it can't still be good, it means it isnt "as good" but its still good. And like I said earlier overall it's a good processor when taking into consideration of its fantastic multi threaded performance and upgrade path.

It's like comparing the 5850 with a 6850. Just because the 5850 is faster it doesnt mean the 6850 still isnt "good". They are both good.

 

[Makaveli]

Just because the Bulldozer isnt consistantly faster than it's replacement in singlethreaded activities doesnt mean it can't still be good, it means it isnt "as good" but its still good. And like I said earlier overall it's a good processor when taking into consideration of its fantastic multi threaded performance and upgrade path.

It's like comparing the 5850 with a 6850. Just because the 5850 is faster it doesnt mean the 6850 still isnt "good". They are both good.

The 6850 is not a replacement for the 5850 in AMD's lineup that would be the 6950 which is faster than the 5850.

if the 6970 was slower than the 5870 everyone would say it was a failure so why does bulldozer get a pass?

 

[Dent1]

The 6850 is not a replacement for the 5850 in AMD's lineup that would be the 6950 which is faster than the 5850.

if the 6970 was slower than the 5870 everyone would say it was a failure so why does bulldozer get a pass?

My argument was according to Wizzards review. The Bulldozer performs on par with the i7 Nehalem. If Bulldozers single threaded performance is so subpar, why does the Nehalem get a pass? Sure its older but I bet 7/10 of TPU members still own a i7 920 or had one, and they wouldn't classify it as a shitty performer at all.

It's not that I'm asking for a pass, I'm just pointing out that people buy whats suitable for them. Some people dont mind average gaming performance if they can encode their videos super quick, doesnt mean the CPU is rubbish, just means its not for you.