AMD Phenom 940 shatters 3DMark World Record - Valentines Day 2009

[magreen]

I seriously LOL'd at the purest silicon thing. Where do they get this stuff? SNL could use some writers like that.

Intel cpus will corrupt your precious bodily fluids! Mothers, grab your children and hold them tight!

 

[Idontcare]

Originally posted by: Zstream

Originally posted by: Idontcare

Originally posted by: eternalone

Originally posted by: Aenslead
Your shattering score just got shattered.

35550 3DMarks on Core i7... what are the ods of that...?

But Phenom 2 was not meant to compete against a Core I7, which makes the argument even more impressive on behalf of the Phenom 2 not to mention the ddr2.

Phenom 2 was not meant to be operated at -196°C either, it was meant to be cooled with the supplied stock HSF. So it makes the clockspeed and resultant suicide-run bench scores all the less impressive...

That makes no sense and you know it. Phenom II was never ever meant to run against I7 in terms of price. If this is your means of replying to his post it... Well nice argument because guess how the I7 score was achieved.

This is a merged post, my comments were a follow-up from a prior OP (which now no longer exists) from the merged thread in which I questioned the value-add of world records generated at these fringe extremes of the system's operation capabilities:

(and yes, I am no more thrilled about the methods in which the i7 scores are being generated, as I said elsewhere in this thread)

Originally posted by: Idontcare

CPU: AMD Engineering Sample

I may be the only person that feels this way, but when I see engineering sample I think "great, who cares?". Of course I tend to think this way when I see that it's cooled with LN2 too. Its not clear to me what the achievement is here. Now if they accomplished this score using off-the-shelf retail available components, including the cooling solution, then I might be convinced there is some value.

But whether they score 35k or 500k in 3dmark06 when they go to otherwise inaccessible extremes to achieve it then for all it matters to me they could be benching some superconducting supercooled alien-tech powered rig on the darkside of the moon, the performance level of their rig means nothing to me personally or as an enthusiast. How excited do we get when the latest o-scopes hit 10GHz? It's a world record clockspeed isn't it? And yet its mostly irrelevant to anyone who doesn't use o-scopes.

Same as a world record 3dmark score from a rig that all but about 10 people on the planet have access to or the knowledge to create with their own hands. Yeah for them :beer:, now lets get back to talking about the kind of performance I can expect to extract from a retail Phenom II rig with commercially available hardware...

Originally posted by: Zstream
The point of benchmarks is to achieve the max out of the hardware.

As I stated in this thread, my personal opinion is that there is essentially no value to the majority of end-users to see the WR scores when non-commercially available means (and CPU's, ES for crying out loud) are employed in order to achieve the "max out of the hardware".

A WR with LN2 means nothing to folks wanting to know if an OC'ed PhII or OC'ed i7 (on air/h2o/phase) is going to be best performance/price for them.

World records in sports aren't allowed for freak conditions (e.g. extreme or atypical conditions), the situations are carefully controlled and yes restricted.

You can't claim the 100m dash world record if you got a 10mph tailwind on your back, for a reason as such a WR would simply be of no-value to the rest of the world that is running without a tailwind.

Can you run faster with a tailwind on your back? Yes. Has someone on the planet actually ran faster than the existing 100m WR holder? Yes they have. Did it count as WR? No it didn't.

If you want the fastest 100m dash, the max the hardware (human body) can give you then strap a rocket to runner and blast them from one end of the race to the other in 0.5 seconds. They probably won't walk again, if not die from the acceleration immediately, but hey its the max the hardware can give you right?

It would also be a WR that no one but save a few individuals would find all that valuable to know about. I'm personally not interested in know how fast a rocket can get a human down the 100m dash. More interested in knowing how fast a human can do it when in their peak physical form but with controlled conditions and the typical running conditions I can expect to find myself in when I go running...

Originally posted by: Zstream
You sir sound just as foolish as the OP.

Your assessment of my mental faculty lost all credibility long ago.

But hey way to step up the personal attack :thumbsup:

 

[Zstream]

Originally posted by: Idontcare

Originally posted by: Zstream

Originally posted by: Idontcare

Originally posted by: eternalone

Originally posted by: Aenslead
Your shattering score just got shattered.

35550 3DMarks on Core i7... what are the ods of that...?

But Phenom 2 was not meant to compete against a Core I7, which makes the argument even more impressive on behalf of the Phenom 2 not to mention the ddr2.

Phenom 2 was not meant to be operated at -196°C either, it was meant to be cooled with the supplied stock HSF. So it makes the clockspeed and resultant suicide-run bench scores all the less impressive...

That makes no sense and you know it. Phenom II was never ever meant to run against I7 in terms of price. If this is your means of replying to his post it... Well nice argument because guess how the I7 score was achieved.

This is a merged post, my comments were a follow-up from a prior OP (which now no longer exists) from the merged thread in which I questioned the value-add of world records generated at these fringe extremes of the system's operation capabilities:

(and yes, I am no more thrilled about the methods in which the i7 scores are being generated, as I said elsewhere in this thread)

Originally posted by: Idontcare

CPU: AMD Engineering Sample

I may be the only person that feels this way, but when I see engineering sample I think "great, who cares?". Of course I tend to think this way when I see that it's cooled with LN2 too. Its not clear to me what the achievement is here. Now if they accomplished this score using off-the-shelf retail available components, including the cooling solution, then I might be convinced there is some value.

But whether they score 35k or 500k in 3dmark06 when they go to otherwise inaccessible extremes to achieve it then for all it matters to me they could be benching some superconducting supercooled alien-tech powered rig on the darkside of the moon, the performance level of their rig means nothing to me personally or as an enthusiast. How excited do we get when the latest o-scopes hit 10GHz? It's a world record clockspeed isn't it? And yet its mostly irrelevant to anyone who doesn't use o-scopes.

Same as a world record 3dmark score from a rig that all but about 10 people on the planet have access to or the knowledge to create with their own hands. Yeah for them :beer:, now lets get back to talking about the kind of performance I can expect to extract from a retail Phenom II rig with commercially available hardware...

Originally posted by: Zstream
The point of benchmarks is to achieve the max out of the hardware.

As I stated in this thread, my personal opinion is that there is essentially no value to the majority of end-users to see the WR scores when non-commercially available means (and CPU's, ES for crying out loud) are employed in order to achieve the "max out of the hardware".

A WR with LN2 means nothing to folks wanting to know if an OC'ed PhII or OC'ed i7 (on air/h2o/phase) is going to be best performance/price for them.

World records in sports aren't allowed for freak conditions (e.g. extreme or atypical conditions), the situations are carefully controlled and yes restricted.

You can't claim the 100m dash world record if you got a 10mph tailwind on your back, for a reason as such a WR would simply be of no-value to the rest of the world that is running without a tailwind.

Can you run faster with a tailwind on your back? Yes. Has someone on the planet actually ran faster than the existing 100m WR holder? Yes they have. Did it count as WR? No it didn't.

If you want the fastest 100m dash, the max the hardware (human body) can give you then strap a rocket to runner and blast them from one end of the race to the other in 0.5 seconds. They probably won't walk again, if not die from the acceleration immediately, but hey its the max the hardware can give you right?

It would also be a WR that no one but save a few individuals would find all that valuable to know about. I'm personally not interested in know how fast a rocket can get a human down the 100m dash. More interested in knowing how fast a human can do it when in their peak physical form but with controlled conditions and the typical running conditions I can expect to find myself in when I go running...

Originally posted by: Zstream
You sir sound just as foolish as the OP.

Your assessment of my mental faculty lost all credibility long ago.

But hey way to step up the personal attack :thumbsup:

It must be over your head to understand so let me repeat again. The point of a benchmark is to get the max out of the hardware. They did not achieve this goal to make you feel better or happy. They did the benchmark to break a world record.

Your analogy can not be used in terms of computers. Nice try though as you struggle for a point. I should have expected this typical response.

 

[Flipped Gazelle]

Originally posted by: Zstream
It must be over your head to understand so let me repeat again. The point of a benchmark is to get the max out of the hardware. They did not achieve this goal to make you feel better or happy. They did the benchmark to break a world record.

Actually, the point of a benchmark is to establish a standard by which performance can be measured, and this can be useful to ascertain that the system is running properly.

 

[Idontcare]

Originally posted by: Zstream

It must be over your head to understand so let me repeat again. The point of a benchmark is to get the max out of the hardware. They did not achieve this goal to make you feel better or happy. They did the benchmark to break a world record.

You are arbitrarily defining the purpose of a benchmark. Then with your arbitrary definition you are attacking me as being inferior in some manner because I do not agree with your arbitrary definition.

benchmark [bench-mahrk]

3. Computers. an established point of reference against which computers or programs can be measured in tests comparing their performance, reliability, etc.

http://dictionary.reference.com/browse/benchmark

I don't see anything about current LN2 cooling implementations to suggest they are a reference point.

I can order third-party air-cooling HSF's, water-cooling gear, and vaporphase equipment which are all off the shelf and available to everyone within shipping distance.

I can't order standardized commercially-produced LN2 computer cooling gear from Newegg or any other retailer.

Originally posted by: Zstream

Your analogy can not be used in terms of computers. Nice try though as you struggle for a point. I should have expected this typical response.

I can make any analogy I want, it is up to you to prove through facts and logic why such an analogy is invalid if in fact it is.

What is typical of your post is that you don't like the analogy I draw because it makes a point that you can't disprove and thus you attack me personally out of desperation and frustration.

This is quite characteristic of a troll. Thanks for outing yourself, again, but keep it up and I really have no qualms about requesting your posting behavior be reviewed by the proper authorities.

 

[Zstream]

Originally posted by: Idontcare

Originally posted by: Zstream

It must be over your head to understand so let me repeat again. The point of a benchmark is to get the max out of the hardware. They did not achieve this goal to make you feel better or happy. They did the benchmark to break a world record.

You are arbitrarily defining the purpose of a benchmark. Then with your arbitrary definition you are attacking me as being inferior in some manner because I do not agree with your arbitrary definition.

benchmark [bench-mahrk]

3. Computers. an established point of reference against which computers or programs can be measured in tests comparing their performance, reliability, etc.

http://dictionary.reference.com/browse/benchmark

I don't see anything about current LN2 cooling implementations to suggest they are a reference point.

I can order third-party air-cooling HSF's, water-cooling gear, and vaporphase equipment which are all off the shelf and available to everyone within shipping distance.

I can't order standardized commercially-produced LN2 computer cooling gear from Newegg or any other retailer.

Originally posted by: Zstream

Your analogy can not be used in terms of computers. Nice try though as you struggle for a point. I should have expected this typical response.

I can make any analogy I want, it is up to you to prove through facts and logic why such an analogy is invalid if in fact it is.

What is typical of your post is that you don't like the analogy I draw because it makes a point that you can't disprove and thus you attack me personally out of desperation and frustration.

This is quite characteristic of a troll. Thanks for outing yourself, again, but keep it up and I really have no qualms about requesting your posting behavior be reviewed by the proper authorities.

Why is the analogy a horrible one? Well lets see...

We are comparing machines vs human beings. What do you think all the land speed records are for? What do you think about the motorcycle, airplane, boat, car speed records.... They use any means outside of cheating to beat the record. Same goes for the 3dmark.

Your analogy of human beings with a tailwind makes no sense. Intel & AMD both use whatever is necessary. They BOTH use Ln2, they both use water, they both use extreme measures to beat the record. They BOTH used picked CPU's... Do I need to go on?

 

[Keysplayr]

ZStream:

"You sir sound just as foolish as the OP."
"It must be over your head to understand so let me repeat again."
"Nice try though as you struggle for a point. I should have expected this typical response."

Feel free to stop the antagonistic posts anytime. Talk about the subject matter until your hearts content, but do not under any circumstances get personal. If you feel the need to insult someone, do not post it.

Anandtech Moderator - Keysplayr2003

 

[SlowSpyder]

My PhII must use dirty, non-pure silicon mined from Mordor. I need 1.5+ volts to hit 3.7GHz. But then again I also have a SB600 that is no doubt holding me back. The SB600's silicon is so dirty, well, it's not even silicon as we know it. Pretty sure it's made from packed sand.

 

[Cookie Monster]

I agree. Your SB600 is holding back the pure siliconess of your PHII.

 

[RallyMaster]

Eww....Mordor silicon is the worst kind. It's almost as bad as Saruman's Uru-k-hai.

 

[ajaidevsingh]

My my actually this silicon is not from Mordor its from Msrdsr "Lost in Middle-earth" which is better silicon and much better than Saruman's Uru-k-hai.

So much is the difference between the two silicons they have different ID's 940 and 945....!! :->

 

[DrMrLordX]

Originally posted by: fusion238
To the two preceding posters, since most world records of any kind are accomplished
utilizing specialized equipment it is to be expected. But what everyone in the world
can infer that the new AMD Phenoms are capable of very good overclocks using very
inexpensive cooling apparatus.

Nothing about a suicide run under LN2 suggests that Phenom IIs are capable of very good overclocks using inexpensive cooling. It suggests that Phenom IIs are capable of very good overclocks under extreme cooling, which is nice since so many chips from AMD have been cold bugged in the past.

If you are trying to infer that future revisions of Phenom II will reach the clock speeds being demonstrated under extreme cooling today, let us all hope that, for AMD's sake, such an event does not occur. Even K8 derivatives, which stayed on the market in various forms as AMD's flagship chip for far too long for them to be competitive, seldom reached stock or air-overclocked speeds equal to those speeds which were achieved with supercooled K8s. Or, to put a finer point on it, Brisbanes did not ship at stock speeds or air-overclock to speeds equal to WR overclocks on Mancesters or Toledos.

In order for AMD to survive, they are going to need to release new CPU revisions before they get to the point that they are forced to market rehashed, die-shrunk Phenom IIs 1-2 years after their release as their flagship products. We don't want to see the AM2/Windsor debacle happen to them again.

There were P4s clocked in excess of 8 ghz. Remember what happened to Intel's plans to ship a P4 with a stock clockspeed of 10 ghz?

WR overclocks don't mean much in the real world.

What we know about Phenom IIs is that, on good air, they're good for about 3.6 to 4.0 ghz using commercially-available silicon. That's nothing to sneeze at, but you can do as well (or better) with Core 2 Quads that have been available for months. Phenom II has the edge when it comes to price 'cuz they're cheap, arguably a tad cheaper than a similar Core 2 Quad system, though it's pretty close.

Phenom II seems to have an edge when you AREN'T overclocking when it comes to price/performance, at least when comparing X4s to Core 2 Quads.

 

[dennilfloss]

Nice video of the deed.

http://www.youtube.com/watch?v=gwkzY8a8aFs&fmt=22

 

[Idontcare]

Originally posted by: dennilfloss
Nice video of the deed.

http://www.youtube.com/watch?v=gwkzY8a8aFs&fmt=22

mmmm yummy viral marketing :thumbsup: (the vid, not the poster)

Not that I mind, it was entertaining to say the least.

Too much with the shaky camera thing though, cloverfield takes on Finland I guess.

 

[Duvie]

The bottomline is....

AMD once again, like they have for the past few years, comes to the dance late with a lesser product then is available at Intel. I dont care if it isn't suppose to compare to the i7 core....It came out afterwards correct?

The phenom II is a big step compared to phenom I for ocing....FACT

The tri core chips are a great idea that Intel should copy....FACT...sidenote maybe this ia a hint of their manufacturing yields

The phenom II is the greatest thing since sliced bread....opinion and viral marketing from the OP

 

[daw123]

Originally posted by: Duvie
The bottomline is....

AMD once again, like they have for the past few years, comes to the dance late with a lesser product then is available at Intel. I dont care if it isn't suppose to compare to the i7 core....It came out afterwards correct?

The phenom II is a big step compared to phenom I for ocing....FACT

The tri core chips are a great idea that Intel should copy....FACT...sidenote maybe this ia a hint of their manufacturing yields

The phenom II is the greatest thing since sliced bread....opinion and viral marketing from the OP

Nice summary. :thumbsup:

 

[Flipped Gazelle]

Originally posted by: Duvie
The bottomline is....

AMD once again, like they have for the past few years, comes to the dance late with a lesser product then is available at Intel. I dont care if it isn't suppose to compare to the i7 core....It came out afterwards correct?

The phenom II is a big step compared to phenom I for ocing....FACT

The tri core chips are a great idea that Intel should copy....FACT...sidenote maybe this ia a hint of their manufacturing yields

The phenom II is the greatest thing since sliced bread....opinion and viral marketing from the OP

I always find these "bottomline" posts amusing, as though the poster were the "Ultimate Authority".

AMD once again, like they have for the past few years, comes to the dance late with a lesser product then is available at Intel. I dont care if it isn't suppose to compare to the i7 core....It came out afterwards correct?

So, by this logic, do you compare, say, a 2008 Toyota Camry with a 2007 Mercedes-Benz S-Class? After all, the Camry came out later...

The tri core chips are a great idea that Intel should copy....FACT...sidenote maybe this ia a hint of their manufacturing yields

1st part is an opinion, not a fact, 2nd part is a guess - with which I happen to agree.

The phenom II is the greatest thing since sliced bread....opinion and viral marketing from the OP

Yup. Notice he hasn't posted in a few days.

 

[dennilfloss]

The Phenom II 940 BE will satisfy my needs at a price that suits my budget: FACT.

 

[fusion238]

World records that involve speed are noteworthy and when you have competition between
AMD's midstream priced CPUs against formidable Intel's highest priced CPUs, it is a great achievement. It is liking being witness to a Mazda RX8 with nitrous oxide outperforming a Jaguar XJ220 twin turbo supercar!

Just as formerly exotic equipment like turbochargers and superchargers were not mainstream, future types of cooling techniques and technologies (including in the CPU)
will bring world record performance to the masses.

If you don't stop this obvious marketing ploy, you will be banned.
Every single post you have made is a marketing post. Not that you own one, and are giving feedback about it, or in reference to an article. (Except this one post referencing an article)

Markfw900 Anandtech Moderator

 

[SlowSpyder]

Originally posted by: fusion238

World records that involve speed are noteworthy and when you have competition between
AMD's midstream priced CPUs against formidable Intel's highest priced CPUs, it is a great achievement. It is liking being witness to a Mazda RX8 with nitrous oxide outperforming a Jaguar XJ220 twin turbo supercar!

Well, I still think you're just a fanboy, and 99.9% of us could care less about these types of suicide bench runs, but you do have a point there. But, when you take a chip, any chip from a low to high end, and you apply that level of exotic cooling/oc methods I don't think it matters anymore what the price of the chip was to start with.

 

[VirtualLarry]

Originally posted by: fusion238

World records that involve speed are noteworthy and when you have competition between
AMD's midstream priced CPUs against formidable Intel's highest priced CPUs

AMD's chips are only "midstream" (mainstream?) priced, because THEY HAVE NO HIGH-END chips.

Face it, the Phenom II 940 IS AMD's highest-end CPU that they produce right now. So attempting to portray this as AMD's mid-grade line against Intel's high-end line is disengenous.

 

[Idontcare]

Originally posted by: fusion238

World records that involve speed are noteworthy and when you have competition between
AMD's midstream priced CPUs against formidable Intel's highest priced CPUs, it is a great achievement. It is liking being witness to a Mazda RX8 with nitrous oxide outperforming a Jaguar XJ220 twin turbo supercar!

Just as formerly exotic equipment like turbochargers and superchargers were not mainstream, future types of cooling techniques and technologies (including in the CPU)
will bring world record performance to the masses.

While your examples are not incorrect, they actually serve the purpose of proving the point others have stated here.

A mazda rx8 w/nitrous beating an XJ220 w/twin-turbo is entertaining to people who are entertained by such things, but it means dick about the performance of the rx8 and xj220 that consumers buy and use.

Meaning the Rx8 w/nitrous vs. XJ220 twin-turbo is pointless when it comes to communicating anything of value to would-be consumers and I'd argue the competition itself is pointless to the manufacturers bottom line.

AMD is not going to sell more 940's because some dudes in Finland overclocked with LN2 and bagged a WR for 6 hrs. Neither is Intel for bagging the WR right after.

In the meantime me knowing that a X4 940 under LN2 competes with an i7 under LN2 means dick as far as me knowing the price/performance I can expect to extract from either an i7 or an X4 940.

These WR stunts are marketing BS thru and thru. Sponsored OC'ers, canned graphics for the youtube vids, viral marketers spamming the forums. It's all BS that adds no value to the consumers who actually pay the bills by buying these things and running them on air.

 

[RallyMaster]

Hah, "competition between AMD's midstream priced CPUs formidable Intel's highest priced CPUs"--I laughed at the RX-8 vs XJ220 comparison. When it's all said and done and you factor in the cooling and time needed to take a 940BE up to Intel's highest end in performance, you'll end up paying about the same anyway. Overclockers don't account for much of the population because honestly, do your parents overclock? Do your grandparents overclock? Would/did they ever? Maybe back in the days of the "turbo button," the answer would have been "Sí, señor." Not the case now, is it with HP/Dell locking down all the BIOS settings and whatnot?

Oh, let's look back on that RX-8 vs XJ220 comparison. If you have an RX-8, you have a Mazda. Period. People will treat it like a Mazda no matter how much you say "It's got $40,000 worth of mods in it and produces 550 bhp." However, if you have a Jaguar XJ220, people will go "That right there, is a supercar and a timeless classic." Of course, you pay a shiny penny for that XJ220 and with the purchase not only comes the supercar performance, but the prestige of owning a car with deep history and a genuine "wow" factor.

That's not the case with owning an Intel chip vs an AMD chip. IMO: Owning an AMD chip says you want to save money while getting a decent amount of performance on the side. Owning an Intel chip means you want performance while spending as much as the AMD fellow. In the end, the difference between the two platforms in terms of pricing for the performance is pretty minimal. I've made this comment at Overclock.net (which is overrun by AMD aficionados, by the way) a couple weeks back. I've actually had some of them say "OMG UR SUCH AN INTEL FANBOIII, GTFO." Well news to you, AMD folks, competition works via fair pricing (whatever's deemed fair for the market) and as far as the current situation is concerned, pricing is so balanced, I had a hard time deciding what to get before I made the jump to quad. The "AMD is cheaper" argument is no longer valid.

RM out.

 

[Markfw]

Originally posted by: RallyMaster
Hah, "competition between AMD's midstream priced CPUs formidable Intel's highest priced CPUs"--I laughed at the RX-8 vs XJ220 comparison. When it's all said and done and you factor in the cooling and time needed to take a 940BE up to Intel's highest end in performance, you'll end up paying about the same anyway. Overclockers don't account for much of the population because honestly, do your parents overclock? Do your grandparents overclock? Would/did they ever? Maybe back in the days of the "turbo button," the answer would have been "Sí, señor." Not the case now, is it with HP/Dell locking down all the BIOS settings and whatnot?

Oh, let's look back on that RX-8 vs XJ220 comparison. If you have an RX-8, you have a Mazda. Period. People will treat it like a Mazda no matter how much you say "It's got $40,000 worth of mods in it and produces 550 bhp." However, if you have a Jaguar XJ220, people will go "That right there, is a supercar and a timeless classic." Of course, you pay a shiny penny for that XJ220 and with the purchase not only comes the supercar performance, but the prestige of owning a car with deep history and a genuine "wow" factor.

That's not the case with owning an Intel chip vs an AMD chip. IMO: Owning an AMD chip says you want to save money while getting a decent amount of performance on the side. Owning an Intel chip means you want performance while spending as much as the AMD fellow. In the end, the difference between the two platforms in terms of pricing for the performance is pretty minimal. I've made this comment at Overclock.net (which is overrun by AMD aficionados, by the way) a couple weeks back. I've actually had some of them say "OMG UR SUCH AN INTEL FANBOIII, GTFO." Well news to you, AMD folks, competition works via fair pricing (whatever's deemed fair for the market) and as far as the current situation is concerned, pricing is so balanced, I had a hard time deciding what to get before I made the jump to quad. The "AMD is cheaper" argument is no longer valid.

RM out.

I guess I don't see your point. First I have an RX-8, and it owns every car under $100,000 in cornering, at Portland speedway at least (and around the country around here). Second, I say the AMD chips just don;t have it. Example, take 2 $200 chips, one from each company, overclock the best you can on hardware that costs the same. Who wins ? Intel. Who wins at the same clock speed ? Intel...

See sig for details on all of the above....

 

[Flipped Gazelle]

Originally posted by: Markfw900
Second, I say the AMD chips just don;t have it. Example, take 2 $200 chips, one from each company, overclock the best you can on hardware that costs the same. Who wins ? Intel. Who wins at the same clock speed ? Intel...

See sig for details on all of the above....

Who wins at stock speed at the same price point?

It seems that X3's do better at their price point than X4's do at theirs.

If you compare the PhII X3 with a comparably priced C2D, there will be some situations where the X3's 3rd core gives it a good leg up.