With the current rate of Intel CPU performance increases, could AMD be catching up?

[strata8]

I also absolutely agree, that the A10-5800K is a MUCH better comparison to the 3770K, than the 8350. :awe:

You use sarcasm very liberally in your arguments. If your goal is to come off as condescending you're doing a very good job of it but otherwise I'd tone it down a little.

 

[SomeoneSimple]

That's okay. I just emphasize how silly Fjodr's thoughts are, saying how Intel has no headroom for higher clocks, while AMD's current "high-end" CPU performs worse in nearly every benchmark while drawing over 60% more power doing so. The fact that Intel has little headroom might, or might not be true, but in any case, currently AMD has even less headroom for higher clocks or TDP.

Also, insisting on comparing the 3770K with the low-performance A10-5800 instead of the power-munching FX8350, just because the 3770K has an on-die IGP, is just laughable, and won't be taken seriously by anyone.

 

[strata8]

That's okay. I just emphasize how silly Fjodr's thoughts are, saying how Intel has no headroom for higher clocks, while AMD's current "high-end" CPU performs worse in nearly every benchmark while drawing over 60% more power doing so. The fact that Intel has little headroom might, or might not be true, but in any case, currently AMD has even less headroom for higher clocks or TDP.

Also, insisting on comparing the 3770K with the low-performance A10-5800 instead of the power-munching FX8350, just because the 3770K has an on-die IGP, is just laughable, and won't be taken seriously by anyone.

I'm pretty amazed at that 168W TDP figure though, and surprised the stock cooler could even handle that.

 

[mrmt]

The FX8350 has 125 W TDP

It hasn't. MSI went on the record to state that FX8350 is 140W.

 

[Idontcare]

I'm pretty amazed at that 168W TDP figure though, and surprised the stock cooler could even handle that.

There is reason the FX8350 stock cooler comes with heatpipes and is substantially beefier than the AMD stock HSF that came with their past 140W TDP CPUs.

 

[Idontcare]

It hasn't. MSI went on the record to state that FX8350 is 140W.

They said it was at least that much, that they set the throttling threshold to 140W.

You should only experience the 255C throttle trip with MSI if your CPU is attempting to pull more than 140W through the mobo.

 

[guskline]

IDC has done extensive testing on the 8350 and though AMD "might" claim 125W TDP I think it is higher. I have an upper end 990FX mb (Asus Sabertooth) so higher TDP doesn't affect me. However, there really isn't much headroom for OCing the 8350 (I'm at 4.6Ghz stable and it still sucks a lot of power when OC'd ( not as much as a Bulldozer but a lot.) For example see my sig. My 3770k is OC' d to 4.4Ghz and has 2 GTX670s in SLI while the FX8350 is OC'd to 4.6Ghz and has a single GTX680.

When I run Intel Burn test on each the 3770k box maxes at @220W while the 8350 is at 390W! Think about it, 2 video cards in the 3770 and 1 in the 8350 yet it draws almost double the power!

 

[Vernon593]

When did we have huge IPC increases from uarch changes?

 

[mrmt]

They said it was at least that much, that they set the throttling threshold to 140W.

You should only experience the 255C throttle trip with MSI if your CPU is attempting to pull more than 140W through the mobo.

Oops... :whiste:

 

[Idontcare]

When did we have huge IPC increases from uarch changes? <Hidden image removed -DrPizza>

For Intel the major IPC bumps were seen at:

386 SX -> 486 DX

486 DX -> Pentium

Netburst -> Core 2 Duo


For AMD it was less pronounced but still there:

K6-2 -> K7

K7 -> K8

?? When else? K5 to K6 was huge but only because the clockspeed nearly tripled, not because IPC made a big jump. The rest were rather mild gains made year over year, adding up to lots of IPC but no major transitions.

 

[Homeles]

The discussion was whether Intel "easily could adjust the frequency upwards on existing CPUs to gain a substantial performance increase with few drawbacks".

If it would mean Intel would go from 77W->125W on a 3770K for a mere 18% performance increase then I guess it's pretty obvious that it's not a reasonable way for Intel to easily increase their CPU performance. I.e. they do not have much headroom to increase the frequency without taking a serious TDP penalty.

It's actually pretty obvious that Intel has hit a frequency wall by looking at the latest CPU releases. SB (2600K) -> IB (3770K) has only brought 100 MHz frequency increase over the 2.5 years that passed between those two CPU releases. It's because Intel knows that increasing the frequency of their current CPUs means a too high TDP penalty to be worth considering (and also because they focus on lowering TDP).

If it really would be easy for Intel to increase the CPU frequency with few drawbacks, don't you think they already would have done so?

Why is this so difficult for you to grasp? Let me ask again: do you think Intel is just going to let AMD waltz in and take the performance crown?

Also, you think 18% is a small number? Lol.

 

[Sleepingforest]

The discussion was whether Intel "easily could adjust the frequency upwards on existing CPUs to gain a substantial performance increase with few drawbacks".

If it would mean Intel would go from 77W->125W on a 3770K for a mere 18% performance increase then I guess it's pretty obvious that it's not a reasonable way for Intel to easily increase their CPU performance. I.e. they do not have much headroom to increase the frequency without taking a serious TDP penalty.

It's actually pretty obvious that Intel has hit a frequency wall by looking at the latest CPU releases. SB (2600K) -> IB (3770K) has only brought 100 MHz frequency increase over the 2.5 years that passed between those two CPU releases. It's because Intel knows that increasing the frequency of their current CPUs means a too high TDP penalty to be worth considering (and also because they focus on lowering TDP).

If it really would be easy for Intel to increase the CPU frequency with few drawbacks, don't you think they already would have done so?

This is not true at all. Actually look at the graph's points. At 125W, the i7-3770K is at 4.6GHz, which is a 31% increase. Yes, it will require a big cooler--but not any more so than the 8350. Yes, the TDP will increase. But it is still lower than the 8350's true power consumption. People think it will cannibalize the socket 2011 sales--untrue, because they offer more cores, and can advertise lower TDP, or bump up the clock speed because of the thermal/voltage headroom, or even better, they can make a new product line which is simply existing chips clocked higher with a higher TDP.

 

[Phynaz]

At 140W TDP they will need a WaterCooling in order to keep the TJmax at 105C.

Then why doesn't the FX8350 need water cooling? Since we know the 125w TDP figure is a lie, does AMD somehow have a grasp of physics that Intel doesn't?

Or there's the other explanation...

 

[AtenRa]

This is not true at all. Actually look at the graph's points. At 125W, the i7-3770K is at 4.6GHz, which is a 31% increase. Yes, it will require a big cooler--but not any more so than the 8350. Yes, the TDP will increase. But it is still lower than the 8350's true power consumption. People think it will cannibalize the socket 2011 sales--untrue, because they offer more cores, and can advertise lower TDP, or bump up the clock speed because of the thermal/voltage headroom, or even better, they can make a new product line which is simply existing chips clocked higher with a higher TDP.

- First of all the TDP is not power consumption.

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".

-Secondly, just because the FX8350 can use that Heat-Sink Fan for 125W TDP it doesnt mean that the Core i7 3770K can use it for 125W TDP. Even if they both had the same TJmax, they still have completely different power maps, the FX8350 must have significant lower temperature hotspots and a bigger die.

-The 2011 CPUs are already 130W(up to 150W TDP for the XEONs) not lower. Also the TJmax is 100c, way higher than the FX8350 and 5c lower than IvyBridge 1155.

77W TDP Intel HSF

130W TDP Intel HSF

Intel charge you $20.00 extra for the HSF if you buy a 130W TDP CPU(Core i7 3820, 3930K etc).
I dont believe they would like to put that extra cost in to the current high volume 1155 socket CPUs and lower their margins and profits.

 

[Homeles]

I dont believe they would like to put that extra cost in to the current high volume 1155 socket CPUs and lower their margins and profits.

This discussion is not about what Intel would like to do, but what they're capable of doing. You, like Fjodor2001, have missed this rather obvious and critical point.

 

[Phynaz]

This is not true at all. Actually look at the graph's points. At 125W, the i7-3770K is at 4.6GHz, which is a 31% increase. Yes, it will require a big cooler--but not any more so than the 8350. Yes, the TDP will increase. But it is still lower than the 8350's true power consumption. People think it will cannibalize the socket 2011 sales--untrue, because they offer more cores, and can advertise lower TDP, or bump up the clock speed because of the thermal/voltage headroom, or even better, they can make a new product line which is simply existing chips clocked higher with a higher TDP.

And there's also binning. Intel manufactures a little less than 1M CPU's a day. If just 5% of those can be cherry picked for lower TDP to come up with a new speed bin that's enough inventory to support a high end release.

 

[AtenRa]

Then why doesn't the FX8350 need water cooling? Since we know the 125w TDP figure is a lie, does AMD somehow have a grasp of physics that Intel doesn't?

Or there's the other explanation...

The 125W TDP is not a lie, again for the millionth time the TDP is not the CPU power consumption.

As to why the FX8350 doesnt need watercooling,

The Core topography (Module design and placement on the die, Cache placement and power map) is completely different than the Quad Core Ivybridge die.
Because it is on 32nm vs 22nm (Ivy) the hotspot temperatures should be way lower, that makes the HSF design simpler/smaller.
Also IvyBridge is not soldered to the Heat spreader like FX8350 and it effects the HSF design and size.

To conclude, in order for Intel to release a 125W 1155 die, they will need to solder the die to the heat spreader and install a huge HSF or WC in order to keep it of reaching the 105c TJmax.

 

[Homeles]

Also IvyBridge is not soldered to the Heat spreader like FX8350 and it effects the HSF design and size.

Even though it's using the same cooler as the soldered Sandy Bridge? Good one.

 

[AtenRa]

Even though it's using the same cooler as the soldered Sandy Bridge? Good one.

and thats why it operates at HIGHER temperatures than the 32nm SandyBridge. :whiste:

edit: Thank you for remind us that, SB is 95W TDP and IV is 77W TDP and yet with the same cooler the 77W TDP CPU operates at higher temperatures. Just imagine what would happen if IV was 125W TDP.

 

[Phynaz]

The 125W TDP is not a lie, again for the millionth time the TDP is not the CPU power consumption.

Okay, then why does AMD not need water cooling for a 150W power consumption but Intel would for 125W?

Die maps? The heat spreader is there for a reason. Do you happen to have temperature maps of both CPU's to prove this point or are you just making this up?

I'm also wondering why the 130W IB Xeon's don't need a soldered IHS and water cooling. Most likely because your posts are full of nonsense.

 

[ShintaiDK]

The 125W TDP is not a lie, again for the millionth time the TDP is not the CPU power consumption.

So you are saying the CPUs can overheat due to underdimentioned cooling? That just makes the FX even worse. We already know they run out of spec and AMD is scamming people in terms of power consumption and putting peoples mobos in danger. But I dont hope they also cheat people on the cooling as well.

 

[Homeles]

and thats why it operates at HIGHER temperatures than the 32nm SandyBridge. :whiste:

edit: Thank you for remind us that, SB is 95W TDP and IV is 77W TDP and yet with the same cooler the 77W TDP CPU operates at higher temperatures. Just imagine what would happen if IV was 125W TDP.

You said it effects (proper word is affects, by the way) the heat sink design. The heat sink design is the same. Therefore you are wrong.

 

[AtenRa]

Okay, then why does AMD not need water cooling for a 150W power consumption but Intel would for 125W?

Die maps? The heat spreader is there for a reason. Do you happen to have temperature maps of both CPU's to prove this point or are you just making this up?

I'm also wondering why the 130W IB Xeon's don't need a soldered IHS and water cooling. Most likely because your posts are full of nonsense.

The FX8350 die size is 315mm2, Ivy is just 160mm2. Not only that but FX module topography(placement) is completely different than Ivy.
Same goes for the 130W TDP 2011 dies, they are 400mm2+ and the core placement is completely different than Ivy 1155 die.

BD/PD power maps are completely different than Ivy and Sandy.

 

[AtenRa]

So you are saying the CPUs can overheat due to underdimentioned cooling? That just makes the FX even worse. We already know they run out of spec and AMD is scamming people in terms of power consumption and putting peoples mobos in danger. But I dont hope they also cheat people on the cooling as well.

Just no,

Let me explain what TDP is for,

The following is from the AMD Phenom II TDP specs.

So, for the CPU to work as it should the HSF should meet the above specs.

That means for a 125W TDP CPU, the HSF at 38c Ambient should be able to keep the CPU temperature at or bellow 61c.

 

[AtenRa]

You said it effects (proper word is affects, by the way) the heat sink design. The heat sink design is the same. Therefore you are wrong.

If i was wrong and they didnt need to redesign the HSF then IvyBridge should have lower operating temperatures than SB.