Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [ChinaDIY] Intel Core i7 4770k "review"
New Posts  All Forums:Forum Nav:

[ChinaDIY] Intel Core i7 4770k "review" - Page 30

post #291 of 298
Quote:
Originally Posted by Brutuz View Post

Exactly, I'm taking this with a grain of salt.
With the additional ~5-10% IPC increase we got from IB + whatever gains we end up seeing from it. (I severely doubt that these are indicative of what final performance will look like.)

We didn't get a 10% increase in IPC, we got better working turbo. If you find a review that has turbo disabled ([H] has one, Sisoft has one with laptops) the difference is much smaller, ranging from a regression to maybe 5%. However there are a few areas where Intel added things to the chip and performance is significantly higher. But regardless there is not enough IPC for IB to justify 4.5ghz max IB clock to 5ghz max SB clock. I would really, really love to see benchmarks between the two when they are at maximum overclock, I have a feeling 3570k and 3770k would lose (and it's probably why there is no IB-E yet, it would be a performance regression).
Quote:
Originally Posted by Brutuz View Post

4.5Ghz vs 5Ghz is a 11% increase, too.

What does that have to do with 5ghz SB to 5ghz Haswell? And lets not forget I'm quite skeptical of the IPC numbers that get thrown around for IB IPC increase as the reviews I've seen with turbo disabled points to IB just getting more IPS due to higher clock speeds from turbo, even though they are advertised as having the same clock speed.
Quote:
Originally Posted by Brutuz View Post

Makes sense, it is meant to be an OCing monster.

5.5 might not be easy if it hits a voltage wall before a thermal wall even on air.

The leaks for SB were actually a lot better, and the same overly optimistic "5ghz+" was thrown around for SB. It was also thrown around for PD too. I think that you should take off 10% or so from what everyone expects to hit, it seems to be the trend. So perhaps 5ghz on solid air and cheap water.
Quote:
Originally Posted by Tsumi View Post

I highly doubt they increased the die size for the sake of improving heat transfer...

I wouldn't be surprised. Intel can sell 90mm^2 i3 for $130 (well, could), and that's desktop. They are competing with 300mm^2+ AMD 2 and 3 modules. They are making massive margins on their chips.

But why? Because the easier it is to cool their chips, the less heatsink needs to go into mobile. The less heatsink, the lower the cost, lower the weight, quieter it is, etc. Making a chip easier to cool would be worth it, and considering Intel's rumored fab usage (read, not near capacity), it would be worth it for them to make it a little bigger to offer better cooling. Do mobile chips even have an IHS?
Tyrant Kuma
(13 items)
 
Starscythe
(13 items)
 
Mobility
(6 items)
 
CPUCPUCPUCPU
Opteron 8431 Opteron 8431 Opteron 8431 Opteron 8431 
MotherboardGraphicsRAMHard Drive
Super Micro h8qme-2+ Sapphire 4870 Samsung DDR2 ECC 667mhz western digital caviar blue 
CoolingCoolingOSPower
FX 8350 stock cooler FX 6300 stock cooler Gentoo Linux 2x Silverstone PSU 
Case
Custom Fabbed Steel Case 
CPUMotherboardGraphicsRAM
AMD FX 8350 gigabyte 990FXA-UD5 7970 Mushkin Enhanced Blackline 16GB 
OSOSMonitorKeyboard
Gentoo Linux Windows 7 x64 Yamakasi Catleap Q270 Corsair K90 
PowerCaseMouseAudio
Silverstone Strider Gold Evolution 850W 80 PLUS... Antec 1200 Logitech G9x Asus Xonar D2X 
Other
XSPC RS360 Raystorm with custom radiator mounts 
CPUGraphicsRAMOS
a4 5000 Radeon HD 8330 8GB G. Skill DDR3 1600 9-9-9 Gentoo Linux 
OSOther
Windows 7 Lenovo Thinkpad x140e 
  hide details  
Reply
Tyrant Kuma
(13 items)
 
Starscythe
(13 items)
 
Mobility
(6 items)
 
CPUCPUCPUCPU
Opteron 8431 Opteron 8431 Opteron 8431 Opteron 8431 
MotherboardGraphicsRAMHard Drive
Super Micro h8qme-2+ Sapphire 4870 Samsung DDR2 ECC 667mhz western digital caviar blue 
CoolingCoolingOSPower
FX 8350 stock cooler FX 6300 stock cooler Gentoo Linux 2x Silverstone PSU 
Case
Custom Fabbed Steel Case 
CPUMotherboardGraphicsRAM
AMD FX 8350 gigabyte 990FXA-UD5 7970 Mushkin Enhanced Blackline 16GB 
OSOSMonitorKeyboard
Gentoo Linux Windows 7 x64 Yamakasi Catleap Q270 Corsair K90 
PowerCaseMouseAudio
Silverstone Strider Gold Evolution 850W 80 PLUS... Antec 1200 Logitech G9x Asus Xonar D2X 
Other
XSPC RS360 Raystorm with custom radiator mounts 
CPUGraphicsRAMOS
a4 5000 Radeon HD 8330 8GB G. Skill DDR3 1600 9-9-9 Gentoo Linux 
OSOther
Windows 7 Lenovo Thinkpad x140e 
  hide details  
Reply
post #292 of 298
Quote:
Originally Posted by sdlvx View Post

We didn't get a 10% increase in IPC, we got better working turbo. If you find a review that has turbo disabled ([H] has one, Sisoft has one with laptops) the difference is much smaller, ranging from a regression to maybe 5%. However there are a few areas where Intel added things to the chip and performance is significantly higher. But regardless there is not enough IPC for IB to justify 4.5ghz max IB clock to 5ghz max SB clock. I would really, really love to see benchmarks between the two when they are at maximum overclock, I have a feeling 3570k and 3770k would lose (and it's probably why there is no IB-E yet, it would be a performance regression).
What does that have to do with 5ghz SB to 5ghz Haswell? And lets not forget I'm quite skeptical of the IPC numbers that get thrown around for IB IPC increase as the reviews I've seen with turbo disabled points to IB just getting more IPS due to higher clock speeds from turbo, even though they are advertised as having the same clock speed.
The leaks for SB were actually a lot better, and the same overly optimistic "5ghz+" was thrown around for SB. It was also thrown around for PD too. I think that you should take off 10% or so from what everyone expects to hit, it seems to be the trend. So perhaps 5ghz on solid air and cheap water.
I wouldn't be surprised. Intel can sell 90mm^2 i3 for $130 (well, could), and that's desktop. They are competing with 300mm^2+ AMD 2 and 3 modules. They are making massive margins on their chips.

But why? Because the easier it is to cool their chips, the less heatsink needs to go into mobile. The less heatsink, the lower the cost, lower the weight, quieter it is, etc. Making a chip easier to cool would be worth it, and considering Intel's rumored fab usage (read, not near capacity), it would be worth it for them to make it a little bigger to offer better cooling. Do mobile chips even have an IHS?

IB had a 0-10% IPC gain with ~7% on average, depending on the program. IB had an improved branch predictor as I recall, and the improvements there depend directly on how often a branch is missed, which is different per program. You cannot base your conclusion off of one program.

SB does not get to 5 ghz on average. More like 4.8.

I have the feeling IB-E was delayed so that the cycle between IB-E and Haswell-E won't be as long. Haswell-E needs DDR4 to be ready before it can be released, and DDR4 isn't due to be ready until late 2014.

Binning a chip for lower power consumption is probably cheaper than making a die larger across an entire lineup for just cooling purposes. Not to mention, a larger die has higher overall power consumption, meaning more overall heat that needs to be dissipated. No matter how you look at it, making a die larger simply for lower temps doesn't make sense.
Millenium Falcon
(24 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 4930k MSI Big Bang Xpower II EVGA GTX 690 Patriot Viper II Sector 7 
Hard DriveHard DriveHard DriveHard Drive
OCZ Deneva 2 Corsair Force 3 Maxtor Western Digital Green 
Optical DriveCoolingCoolingCooling
Samsung BD/DVD-RW Swiftech MCP655 x2 Black Ice GTX 480 Black Ice GTX 280 
CoolingCoolingCoolingCooling
Alphacool Repack Dual D5 Watercool Heatkiller 3.0 Alphacool GTX 690 fullcover Bitspower Big Bang Xpower II fullcover 
OSMonitorKeyboardPower
Windows 8.1 64-bit Professional 3x Dell S2340 Max Keyboard Durandal CoolerMaster V1000 
CaseMouseMouse PadAudio
Azza Genesis 9000B Logitech G700 Roccat Alumic Onkyo HT-S9100THX 
  hide details  
Reply
Millenium Falcon
(24 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 4930k MSI Big Bang Xpower II EVGA GTX 690 Patriot Viper II Sector 7 
Hard DriveHard DriveHard DriveHard Drive
OCZ Deneva 2 Corsair Force 3 Maxtor Western Digital Green 
Optical DriveCoolingCoolingCooling
Samsung BD/DVD-RW Swiftech MCP655 x2 Black Ice GTX 480 Black Ice GTX 280 
CoolingCoolingCoolingCooling
Alphacool Repack Dual D5 Watercool Heatkiller 3.0 Alphacool GTX 690 fullcover Bitspower Big Bang Xpower II fullcover 
OSMonitorKeyboardPower
Windows 8.1 64-bit Professional 3x Dell S2340 Max Keyboard Durandal CoolerMaster V1000 
CaseMouseMouse PadAudio
Azza Genesis 9000B Logitech G700 Roccat Alumic Onkyo HT-S9100THX 
  hide details  
Reply
post #293 of 298
Quote:
Originally Posted by Forceman View Post

Isn't all that extra area just the increased size of the iGPU though?  I think it might help some, but the thermal density of the actual cores should be the same, right?
I've read a couple of datasheets and other PDF's but it isn't clear to me what it is. And if I'm not reading it wrong then I believe that Intel improved the thermal density. Like a couple of other things that are still under NDA.
Quote:
Originally Posted by Tsumi View Post

I highly doubt they increased the die size for the sake of improving heat transfer...
I don't think that's the case, no, but a larger die is said to transfer heat better.

p.s. Seem like not all boards are the same/as good as the others, because I have one here that can't go past 5.2GHz. As in it won't even boot. Whatever I do. No go.
post #294 of 298
The 4770k is looking wonderful, and for everyone interested, there is actually a giveaway going on right now by Linus Tech Tips. Check it out! Hope this helps

http://linustechtips.com/main/topic/20886-intel-free-lan-event-in-dupont-wa-plus-a-chance-to-win-an-unlocked-4th-generation-core-i7/
post #295 of 298
Quote:
Originally Posted by Tsumi View Post

IB had a 0-10% IPC gain with ~7% on average, depending on the program. IB had an improved branch predictor as I recall, and the improvements there depend directly on how often a branch is missed, which is different per program. You cannot base your conclusion off of one program.

SB does not get to 5 ghz on average. More like 4.8.

I have the feeling IB-E was delayed so that the cycle between IB-E and Haswell-E won't be as long. Haswell-E needs DDR4 to be ready before it can be released, and DDR4 isn't due to be ready until late 2014.

Binning a chip for lower power consumption is probably cheaper than making a die larger across an entire lineup for just cooling purposes. Not to mention, a larger die has higher overall power consumption, meaning more overall heat that needs to be dissipated. No matter how you look at it, making a die larger simply for lower temps doesn't make sense.

I posted a picture of a performance regression from SB to IB with turbo disabled and you just literally spouted your touchy feely personal beliefs without citing any sources. I simply claimed that IB's gains are massively over-rated and posted a picture of it doing worse than the previous generation. You are now claiming that it has a 7% IPC increase. Please provide proof.

It is not just [H]orribleOCP either:

If anything IB has a few select instances where it is better clock for clock but overall it doesn't do much. More degradation in performance going from SB to IB. If anything the folks saying IB is 5% or 10% faster per clock because they are looking at reviews with two CPUs with turbo enabled (on different process nodes, no less!) have been drinking a lot of Kool Aide.
Tyrant Kuma
(13 items)
 
Starscythe
(13 items)
 
Mobility
(6 items)
 
CPUCPUCPUCPU
Opteron 8431 Opteron 8431 Opteron 8431 Opteron 8431 
MotherboardGraphicsRAMHard Drive
Super Micro h8qme-2+ Sapphire 4870 Samsung DDR2 ECC 667mhz western digital caviar blue 
CoolingCoolingOSPower
FX 8350 stock cooler FX 6300 stock cooler Gentoo Linux 2x Silverstone PSU 
Case
Custom Fabbed Steel Case 
CPUMotherboardGraphicsRAM
AMD FX 8350 gigabyte 990FXA-UD5 7970 Mushkin Enhanced Blackline 16GB 
OSOSMonitorKeyboard
Gentoo Linux Windows 7 x64 Yamakasi Catleap Q270 Corsair K90 
PowerCaseMouseAudio
Silverstone Strider Gold Evolution 850W 80 PLUS... Antec 1200 Logitech G9x Asus Xonar D2X 
Other
XSPC RS360 Raystorm with custom radiator mounts 
CPUGraphicsRAMOS
a4 5000 Radeon HD 8330 8GB G. Skill DDR3 1600 9-9-9 Gentoo Linux 
OSOther
Windows 7 Lenovo Thinkpad x140e 
  hide details  
Reply
Tyrant Kuma
(13 items)
 
Starscythe
(13 items)
 
Mobility
(6 items)
 
CPUCPUCPUCPU
Opteron 8431 Opteron 8431 Opteron 8431 Opteron 8431 
MotherboardGraphicsRAMHard Drive
Super Micro h8qme-2+ Sapphire 4870 Samsung DDR2 ECC 667mhz western digital caviar blue 
CoolingCoolingOSPower
FX 8350 stock cooler FX 6300 stock cooler Gentoo Linux 2x Silverstone PSU 
Case
Custom Fabbed Steel Case 
CPUMotherboardGraphicsRAM
AMD FX 8350 gigabyte 990FXA-UD5 7970 Mushkin Enhanced Blackline 16GB 
OSOSMonitorKeyboard
Gentoo Linux Windows 7 x64 Yamakasi Catleap Q270 Corsair K90 
PowerCaseMouseAudio
Silverstone Strider Gold Evolution 850W 80 PLUS... Antec 1200 Logitech G9x Asus Xonar D2X 
Other
XSPC RS360 Raystorm with custom radiator mounts 
CPUGraphicsRAMOS
a4 5000 Radeon HD 8330 8GB G. Skill DDR3 1600 9-9-9 Gentoo Linux 
OSOther
Windows 7 Lenovo Thinkpad x140e 
  hide details  
Reply
post #296 of 298
0.6% is well within the margin of error of tests. Like I said, the IPC gains range from 0-10%, depending on the software. 0% is an expected result, as is 10%.

You calling [H]OCP names is not going to refute their test results. Neither is only one benchmark going to prove your point.
Millenium Falcon
(24 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 4930k MSI Big Bang Xpower II EVGA GTX 690 Patriot Viper II Sector 7 
Hard DriveHard DriveHard DriveHard Drive
OCZ Deneva 2 Corsair Force 3 Maxtor Western Digital Green 
Optical DriveCoolingCoolingCooling
Samsung BD/DVD-RW Swiftech MCP655 x2 Black Ice GTX 480 Black Ice GTX 280 
CoolingCoolingCoolingCooling
Alphacool Repack Dual D5 Watercool Heatkiller 3.0 Alphacool GTX 690 fullcover Bitspower Big Bang Xpower II fullcover 
OSMonitorKeyboardPower
Windows 8.1 64-bit Professional 3x Dell S2340 Max Keyboard Durandal CoolerMaster V1000 
CaseMouseMouse PadAudio
Azza Genesis 9000B Logitech G700 Roccat Alumic Onkyo HT-S9100THX 
  hide details  
Reply
Millenium Falcon
(24 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 4930k MSI Big Bang Xpower II EVGA GTX 690 Patriot Viper II Sector 7 
Hard DriveHard DriveHard DriveHard Drive
OCZ Deneva 2 Corsair Force 3 Maxtor Western Digital Green 
Optical DriveCoolingCoolingCooling
Samsung BD/DVD-RW Swiftech MCP655 x2 Black Ice GTX 480 Black Ice GTX 280 
CoolingCoolingCoolingCooling
Alphacool Repack Dual D5 Watercool Heatkiller 3.0 Alphacool GTX 690 fullcover Bitspower Big Bang Xpower II fullcover 
OSMonitorKeyboardPower
Windows 8.1 64-bit Professional 3x Dell S2340 Max Keyboard Durandal CoolerMaster V1000 
CaseMouseMouse PadAudio
Azza Genesis 9000B Logitech G700 Roccat Alumic Onkyo HT-S9100THX 
  hide details  
Reply
post #297 of 298
Quote:
Originally Posted by Pike View Post

Quote:
Originally Posted by Forceman View Post

Isn't all that extra area just the increased size of the iGPU though?  I think it might help some, but the thermal density of the actual cores should be the same, right?
I've read a couple of datasheets and other PDF's but it isn't clear to me what it is. And if I'm not reading it wrong then I believe that Intel improved the thermal density. Like a couple of other things that are still under NDA.
Quote:
Originally Posted by Tsumi View Post

I highly doubt they increased the die size for the sake of improving heat transfer...
I don't think that's the case, no, but a larger die is said to transfer heat better.

p.s. Seem like not all boards are the same/as good as the others, because I have one here that can't go past 5.2GHz. As in it won't even boot. Whatever I do. No go.

Are you cereal bro? Heard you say a lot.. post something. A pic is worth.. well you know.
post #298 of 298
Quote:
Originally Posted by CL3P20 View Post

Quote:
Originally Posted by Pike View Post

Quote:
Originally Posted by Forceman View Post

Isn't all that extra area just the increased size of the iGPU though?  I think it might help some, but the thermal density of the actual cores should be the same, right?
I've read a couple of datasheets and other PDF's but it isn't clear to me what it is. And if I'm not reading it wrong then I believe that Intel improved the thermal density. Like a couple of other things that are still under NDA.
Quote:
Originally Posted by Tsumi View Post

I highly doubt they increased the die size for the sake of improving heat transfer...
I don't think that's the case, no, but a larger die is said to transfer heat better.

p.s. Seem like not all boards are the same/as good as the others, because I have one here that can't go past 5.2GHz. As in it won't even boot. Whatever I do. No go.

Are you cereal bro? Heard you say a lot.. post something. A pic is worth.. well you know.
Fruit Loops....?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Rumors and Unconfirmed Articles
Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [ChinaDIY] Intel Core i7 4770k "review"