General CPU question

drotto25 · 09-12-07, 10:07 AM

I have always wondered how a chip becomes a e6750 vs an e6850 for example. They are based on the same silicone, seem to have the same transistors, etc. So if they are made the same and the basic specs are the same save ghz, why are some given a greater speed and sold for more. This is especially true with the current qx line.

Part 2 to that is why chips that are sold as the exact same chip with the same stepping clock differently. Does anyone know why this is?

JamesXP · 09-12-07, 10:12 AM

Because they manafacture the chips and them certain chips are only stable to E6750 levels not E6850 levels.

(Binning)

drotto25 · 09-12-07, 10:14 AM

But what do they use as their benchmark to determine stability? Especially when as we prove here many can be stable at oc's.

Maverick0984 · 09-12-07, 10:19 AM

Super powerful commercial devices.

JamesXP · 09-12-07, 10:47 AM

I could see intel fabs benching they're thousands of CPUs 24HR with TAT

th3 · 09-12-07, 10:48 AM

One other difference i noticed, in the FPO # you often see the "B" one the E6850s while lower models are mostly "A" (Like in L720Axxx). Also on the first Conroes there was a noticeable difference in OCability between A and B. Read somewhere that A and B silicon came from different US fabs, and the other letters (F and G) were Ireland and Israel. That part might be wrong though, anyway Bs were popular among overclockers a year ago and still are, i jumped on the B wagon myself and got a very nice improvement.

Maverick0984 · 09-12-07, 10:59 AM

It is generally just luck of the draw. My personal opinion is that the whole craze of A vs. B or whatever you are talking about is mostly placebo and a little bit of luck. There will be good and bad chips from every batch in terms of OC ability. Intel only guarantees them at stock settings. We are able to OC so well because they "aim low" on the binning/rating.

diaz · 09-12-07, 12:11 PM

They clock each die to find out which silicon handles what frequency, then they set a multiplier and add the IHS to the substrate. Each batch of silicon has variances so when they test for highest clock, it has little bearing on the rest of the bin's capabilities. So sometimes you could end up with a E6300 that clocks to 3.8, or with a E6600 that wont clock past 3.0.

-D

JamesXP · 09-12-07, 01:39 PM

Isn't A Made in such a place And B made in malasia

Quote me if I'm wrong.

09-12-07, 10:07 AM	Thread Starter #1
drotto25 Member Join Date: Sep 2007 Location: New Jersey	General CPU question I have always wondered how a chip becomes a e6750 vs an e6850 for example. They are based on the same silicone, seem to have the same transistors, etc. So if they are made the same and the basic specs are the same save ghz, why are some given a greater speed and sold for more. This is especially true with the current qx line. Part 2 to that is why chips that are sold as the exact same chip with the same stepping clock differently. Does anyone know why this is?
	QUOTE Thanks

09-12-07, 10:12 AM	#2
JamesXP Epic Fail Guy Join Date: Apr 2007 Location: England	Because they manafacture the chips and them certain chips are only stable to E6750 levels not E6850 levels. (Binning) __________________ Intel i5 3570k @ 4.2GHz 8GB Kingston HyperX DDR 1600Mhz nVidia 8600GT ASRock Z77 Pro4 128GB Samsung 830 SSD Corsair VX450 My Saitek Eclipse III Review
	QUOTE Thanks

09-12-07, 10:14 AM	Thread Starter #3
drotto25 Member Join Date: Sep 2007 Location: New Jersey	But what do they use as their benchmark to determine stability? Especially when as we prove here many can be stable at oc's.
	QUOTE Thanks

09-12-07, 10:19 AM	#4
Maverick0984 Member Join Date: Jul 2007 Location: Folding from Chicago	Super powerful commercial devices. __________________ 3DMark06 - 14,936/XP x86 - 14,172/Vista x64 ASUS P5K Premium WiFi/AP Q6600 @ 3.645 Ghz (405x9) @ 1.425v eVGA Geforce 8800 GTX @ 621Mhz/1512Mhz/1053Mhz 4x2GB G.Skill PK's DDR2-800 @ 405Mhz 1:1 4-4-4-10 2T @ 2.05v 2x150GB Raptor 10,000RPM in RAID 1 w/ ICH9R 7x750GB WD in RAID 5 w/ 3ware 9650SE 8 Port controller 750W PC Power & Cooling
	QUOTE Thanks

09-12-07, 10:47 AM	#5
JamesXP Epic Fail Guy Join Date: Apr 2007 Location: England	I could see intel fabs benching they're thousands of CPUs 24HR with TAT __________________ Intel i5 3570k @ 4.2GHz 8GB Kingston HyperX DDR 1600Mhz nVidia 8600GT ASRock Z77 Pro4 128GB Samsung 830 SSD Corsair VX450 My Saitek Eclipse III Review
	QUOTE Thanks

General CPU question

Internet Services

Website Development

Tech Hardware

09-12-07, 10:48 AM	#6
th3 Member Join Date: Aug 2007	One other difference i noticed, in the FPO # you often see the "B" one the E6850s while lower models are mostly "A" (Like in L720Axxx). Also on the first Conroes there was a noticeable difference in OCability between A and B. Read somewhere that A and B silicon came from different US fabs, and the other letters (F and G) were Ireland and Israel. That part might be wrong though, anyway Bs were popular among overclockers a year ago and still are, i jumped on the B wagon myself and got a very nice improvement.
	QUOTE Thanks

09-12-07, 10:59 AM	#7
Maverick0984 Member Join Date: Jul 2007 Location: Folding from Chicago	It is generally just luck of the draw. My personal opinion is that the whole craze of A vs. B or whatever you are talking about is mostly placebo and a little bit of luck. There will be good and bad chips from every batch in terms of OC ability. Intel only guarantees them at stock settings. We are able to OC so well because they "aim low" on the binning/rating. __________________ 3DMark06 - 14,936/XP x86 - 14,172/Vista x64 ASUS P5K Premium WiFi/AP Q6600 @ 3.645 Ghz (405x9) @ 1.425v eVGA Geforce 8800 GTX @ 621Mhz/1512Mhz/1053Mhz 4x2GB G.Skill PK's DDR2-800 @ 405Mhz 1:1 4-4-4-10 2T @ 2.05v 2x150GB Raptor 10,000RPM in RAID 1 w/ ICH9R 7x750GB WD in RAID 5 w/ 3ware 9650SE 8 Port controller 750W PC Power & Cooling
	QUOTE Thanks

09-12-07, 12:11 PM	#8
diaz Member Join Date: Aug 2007 Location: Canada	They clock each die to find out which silicon handles what frequency, then they set a multiplier and add the IHS to the substrate. Each batch of silicon has variances so when they test for highest clock, it has little bearing on the rest of the bin's capabilities. So sometimes you could end up with a E6300 that clocks to 3.8, or with a E6600 that wont clock past 3.0. -D __________________ \m/ OverClockers mATX L33T Club \m/ >>>JOIN NOW<<< Main Rig i5 3570k EVGA GTX 680 Asus Maximus V Gene mATX Silverstone TJ08E mATX 16G Corsair Vengeance Corsair H80 Silverstone Strider Plus 850 Crucial M4 256G SSD // Seagate 750G 7200.12 Storage Recommended PSU's - True/Tested/**/Intel C2D/Q/X Thermal Designs Explained Heatware Last edited by diaz; 09-12-07 at 12:16 PM.*
	QUOTE Thanks

09-12-07, 01:39 PM	#9
JamesXP Epic Fail Guy Join Date: Apr 2007 Location: England	Isn't A Made in such a place And B made in malasia Quote me if I'm wrong. __________________ Intel i5 3570k @ 4.2GHz 8GB Kingston HyperX DDR 1600Mhz nVidia 8600GT ASRock Z77 Pro4 128GB Samsung 830 SSD Corsair VX450 My Saitek Eclipse III Review
	QUOTE Thanks

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search