• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Sandy Bridge Processors

Thanks for all the info Guyonthecouch. Is there an option in BIOS to turn off the integrated GPU if you're using a video card?
 
Karndog said:
Thanks for all the info Guyonthecouch. Is there an option in BIOS to turn off the integrated GPU if you're using a video card?
Click to expand...

Yes, I believe the CPU disabled the internal video upon detecting a video card installed.

I'd need to test with a board that has integrated video to verify, but I do believe there's an AUTO switch, similar to using a board with onboard video and inserting a PCI-E card.
 
Cool thanks. Yeah i'd hate to think that CPU horsepower was being wasted powering a GPU that isnt even being used when you have a dedicated video card. Cheers.
 
Karndog said:
Cool thanks. Yeah i'd hate to think that CPU horsepower was being wasted powering a GPU that isnt even being used when you have a dedicated video card. Cheers.
Click to expand...

No, that won't be the case at all. The onboard GPU will not draw very much power as it is and when it isn't in use (SLEEP OR STANDYBY) it won't draw any power at all. This would be true of the CPU in an Idle state as well, the GPU would be drawing negligible power.

With a dedicated PCI-E card, the onboard GPU will be disabled. Now, there may be a BIOS release that allows a Hybrid mode (COUNT ON IT) where the onboard GPU is enabled initially and then is turned off in gaming scenarios and handed off to a dedicated GPU. I can totally see that being a feature on a Gigabyte board...
 
TheGuyOnTheCouch said:
No, that won't be the case at all. The onboard GPU will not draw very much power as it is and when it isn't in use (SLEEP OR STANDYBY) it won't draw any power at all. This would be true of the CPU in an Idle state as well, the GPU would be drawing negligible power.

With a dedicated PCI-E card, the onboard GPU will be disabled. Now, there may be a BIOS release that allows a Hybrid mode (COUNT ON IT) where the onboard GPU is enabled initially and then is turned off in gaming scenarios and handed off to a dedicated GPU. I can totally see that being a feature on a Gigabyte board...
Click to expand...

Is Gigabyte also going to provide a hands-off DVI/DP switcher? I'm not saying hybrid mode is a bad idea, but the discrete vs. onboard connector issue needs to be handled first.
 
Surly73 said:
Is Gigabyte also going to provide a hands-off DVI/DP switcher? I'm not saying hybrid mode is a bad idea, but the discrete vs. onboard connector issue needs to be handled first.
Click to expand...

I am sure that would be the easiest route. I am not even 100% sure this will happen, but as a concept, I am sure it's already being looked at by board manufacturers.
 
vengence said:
As mentioned before, the clock generator was moved onto the CPU. That really doesn't tell the whole story though. It's an architectural decision that kills overclocking. Now the argument is it was needed for on die GPU. I'm not sure that's something I'm buying though TBH.
Click to expand...

Why? I think its too easy to lose perspective on these boards. The world of [H], Anand, Toms, XTreme, et al *feels* big, but trust me, to Intel it is massively insignificant. It's a fun diversion.

Give the industry trajectory, these architectural shifts are almost certainly about agility in power consumption and performance per watt. They want greater workload elasticity which means moving more control into the core.

With cloud build outs accelerating, and virtualization becoming a near given in any high density server application, these capabilities are more important than ever.

Remember... the two HIGH revenue areas for Intel are servers and portables. The financial industry, Microsoft, Google and Amazon buy more Intel procs than the entire enthusiast community combined by orders of magnitude that are ridiculous. On the other side are the tens of millions of basic commodity notebooks.

BOTH of these use cases benefit massively from increased power per watt agility in the proc and chipset. If enthusiast overclocking is a sacrifice, no one even notices. Its a blip. A concession to the community will be the unlocked parts and as an earlier poster pointed out, the Intel enthusiast community has been fine paying a bit more.

Folks are looking at this from the wrong direction. Its easy to view it from the direction of the student on a budget who is being shut down by the huge corporation who doesnt want you to "get away with" getting $500 performance out of $100 part. The reality is that Intel is looking at million plus box orders from OEMs and are adjusting their roadmap for emerging trends 2, 3, 10 years out. They really arent sitting in a room figuring out how they can extract an extra $100 from the extremely small number of impoverished enthusiasts who are looking to squeeze maximum performance from the cheapest part. I also distrust corporations reflexively. Ive worked for some of the biggest around for 20 years and have seen enough to know it is wise to. This almost certainly isnt one of those cases though. I have seen this from the inside and decisions like this are about industry trajectory and momentum.

@GuyontheCouch... great info! you're a bolder man than I :)
 
Last edited:
mlambert890 said:
Why? I think its too easy to lose perspective on these boards. The world of [H], Anand, Toms, XTreme, et al *feels* big, but trust me, to Intel it is massively insignificant. It's a fun diversion.

Give the industry trajectory, these architectural shifts are almost certainly about agility in power consumption and performance per watt. They want greater workload elasticity which means moving more control into the core.

With cloud build outs accelerating, and virtualization becoming a near given in any high density server application, these capabilities are more important than ever.

Remember... the two HIGH revenue areas for Intel are servers and portables. The financial industry, Microsoft, Google and Amazon buy more Intel procs than the entire enthusiast community combined by orders of magnitude that are ridiculous. On the other side are the tens of millions of basic commodity notebooks.

BOTH of these use cases benefit massively from increased power per watt agility in the proc and chipset. If enthusiast overclocking is a sacrifice, no one even notices. Its a blip. A concession to the community will be the unlocked parts and as an earlier poster pointed out, the Intel enthusiast community has been fine paying a bit more.

Folks are looking at this from the wrong direction. Its easy to view it from the direction of the student on a budget who is being shut down by the huge corporation who doesnt want you to "get away with" getting $500 performance out of $100 part. The reality is that Intel is looking at million plus box orders from OEMs and are adjusting their roadmap for emerging trends 2, 3, 10 years out. They really arent sitting in a room figuring out how they can extract an extra $100 from the extremely small number of impoverished enthusiasts who are looking to squeeze maximum performance from the cheapest part. I also distrust corporations reflexively. Ive worked for some of the biggest around for 20 years and have seen enough to know it is wise to. This almost certainly isnt one of those cases though. I have seen this from the inside and decisions like this are about industry trajectory and momentum.

@GuyontheCouch... great info! you're a bolder man than I :)
Click to expand...
^Good info... makes a lot of sense..

I'm just wondering if Intel's going to come out with a socket that'll be relevant for as long as LGA775 was... I'm buying a new motherboard soon and I don't want to be stuck with a dead platform in 18 months..
 
mlambert890 said:
Why? I think its too easy to lose perspective on these boards. The world of [H], Anand, Toms, XTreme, et al *feels* big, but trust me, to Intel it is massively insignificant. It's a fun diversion.

Give the industry trajectory, these architectural shifts are almost certainly about agility in power consumption and performance per watt. They want greater workload elasticity which means moving more control into the core.

With cloud build outs accelerating, and virtualization becoming a near given in any high density server application, these capabilities are more important than ever.

Remember... the two HIGH revenue areas for Intel are servers and portables. The financial industry, Microsoft, Google and Amazon buy more Intel procs than the entire enthusiast community combined by orders of magnitude that are ridiculous. On the other side are the tens of millions of basic commodity notebooks.

BOTH of these use cases benefit massively from increased power per watt agility in the proc and chipset. If enthusiast overclocking is a sacrifice, no one even notices. Its a blip. A concession to the community will be the unlocked parts and as an earlier poster pointed out, the Intel enthusiast community has been fine paying a bit more.

Folks are looking at this from the wrong direction. Its easy to view it from the direction of the student on a budget who is being shut down by the huge corporation who doesnt want you to "get away with" getting $500 performance out of $100 part. The reality is that Intel is looking at million plus box orders from OEMs and are adjusting their roadmap for emerging trends 2, 3, 10 years out. They really arent sitting in a room figuring out how they can extract an extra $100 from the extremely small number of impoverished enthusiasts who are looking to squeeze maximum performance from the cheapest part. I also distrust corporations reflexively. Ive worked for some of the biggest around for 20 years and have seen enough to know it is wise to. This almost certainly isnt one of those cases though. I have seen this from the inside and decisions like this are about industry trajectory and momentum.

@GuyontheCouch... great info! you're a bolder man than I :)
Click to expand...

You're argument doesn't hold water. They are still controlling the multiplier from the motherboard. They are still controlling the voltages from the motherboard. And if it's a non-overclocker then the FSB will be on the motherboard and set to 100mhz so absolutely nothing had changed in regards to flexability of downclocking/undervolting.
 
I was going to build an I7 rig for Photoshop and maybe some gaming. With SB being so close I decided to wait on it. I am in need of a video card right now and I'm undecided whether to go with either a Gigabye SOC GTX 460 or GTX 470.

Just curious whether the 1155 board(s) will handle SLI? HW Canucks shows 3 new baords from Gigabyte and mentioned they'll do CF but they weren't sure about SLI.

Jarrod
 
I believe I've read about an 1155 board with the nForce200 chipset in order to operate SLI at 16x16
 
The i5 2500K seems to be the perfect chip for a new gaming rig. Cheap(200-250$?), good overclocking, and should run pretty cool. And I like the fact that if a video card goes bad, you can use the cpu gpu until your new one comes in and not be SOL for 2 weeks.
 
sok0 said:
The i5 2500K seems to be the perfect chip for a new gaming rig. Cheap(200-250$?), good overclocking, and should run pretty cool. And I like the fact that if a video card goes bad, you can use the cpu gpu until your new one comes in and not be SOL for 2 weeks.
Click to expand...
Only if using an Hxx board. The P series boards don't support the GPU (same as current and every Gxxx/Pxxx board in the past). And there are a couple of downsides to the H series, among them a (rumored) lack of memory OC ability (and I for one would like to get DDR3-1600/CL7 RAM and have it run at DDR3-1600).
 
The 2011 is where it's at performance wise.

*power and ground lines to support 40 PCIe v3 lanes
*four DDR-1600 memory channels per socket, dual QPI 8 gigatransfers/sec links
*8-core, 20MB L3 cache Sandy Bridge-based Xeons
 
Mega666 said:
The 2011 is where it's at performance wise.

*power and ground lines to support 40 PCIe v3 lanes
*four DDR-1600 memory channels per socket, dual QPI 8 gigatransfers/sec links
*8-core, 20MB L3 cache Sandy Bridge-based Xeons
Click to expand...

Sandy Bridge E is so sexy.

Great stuff so far.
 
Also shit for turbo and poor Ghz/$ to begin with, same as s1366. They'll be good chips, but the only people that really need them, *really* need them. And they're mediocre for the rest of us.
 
GLSauron said:
Also shit for turbo and poor Ghz/$ to begin with, same as s1366. They'll be good chips, but the only people that really need them, *really* need them. And they're mediocre for the rest of us.
Click to expand...

We should wait until price information is released before making any assumptions.
 
TheGuyOnTheCouch said:
We should wait until price information is released before making any assumptions.
Click to expand...
The $562 bracket for s1366 is the i7-960 (3.2-->3.46Ghz). The same bracket for s1155 will be i7-2600 (3.4-->3.8Ghz). Those'll change, but I question by how much.

edit: thats a good point. The die is the same, but I shouldn't have been talking about Xeon pricing.
 
Last edited:
LOL the Dual QPI/20MB L3 Cache/ Quad Channel DDR3 will be the 2000-3000 dollar chips.

The ones that will resemble the I7 920-990x will be the ones with 1x QPI/15MB L3 Cache/Triple Channel DDR3.

Im not so sure if i should wait for a 6 core Sandy Bridge-E or not. Not looking like much advantage over a I7 6 core unless i want to buy a server board and a 1400 dollar Xeon just to get the new stuff.

Still on the fence but im not liking SB-E so far.
 
I think I will stick with the i7 for a few more years and then move on to the new Intel chips. The i7 tears through everything.
 
So i5-2500K will be 3.3GHz, 6MB, 4 core, 4 thread, and 95W TDP.
i7-2600K will be 3.4 GHz, 8MB, 4 core, 8 thread, and 95W TDP.

I'm thinking 2500K is the way to go in this respect. I don't see the extra 2MB, .1GHz, or multithreading as really being beneficial. Anyone want to correct me on this notion?
 
haxxorboi said:
So i5-2500K will be 3.3GHz, 6MB, 4 core, 4 thread, and 95W TDP.
i7-2600K will be 3.4 GHz, 8MB, 4 core, 8 thread, and 95W TDP.

I'm thinking 2500K is the way to go in this respect. I don't see the extra 2MB, .1GHz, or multithreading as really being beneficial. Anyone want to correct me on this notion?
Click to expand...

It's just like the current I5 vs. I7 argument. Except there is extra cache here.
 
haxxorboi said:
So i5-2500K will be 3.3GHz, 6MB, 4 core, 4 thread, and 95W TDP.
i7-2600K will be 3.4 GHz, 8MB, 4 core, 8 thread, and 95W TDP.

I'm thinking 2500K is the way to go in this respect. I don't see the extra 2MB, .1GHz, or multithreading as really being beneficial. Anyone want to correct me on this notion?
Click to expand...

Really would depend on the pricing.
 
Ehren8879 said:
I may pony up an extra $100 - 125 for the extra cache and HT features, but if it's a $500+ chip then no way.
Click to expand...

The extra cache and HT might be worth a lot more to somebody, you never know.

Personally, the I5 Sandy Bridge is going to be a great value segment, but I need to see pricing on the I7's to make any determination of value vs.
 
Looking forward to new i7-2600k. I have no objection to paying for absolutely the best. Can't wait for 6-core Sandy Bridge also. They look like a great CPU. Still no word when Sandy Bridge Xeon & 2MP cpus will be ready/realeased. So my workstations are stable for now.

Have 1 975X and 2 980X PCs and a i7-875k thrown in for good measure. I do overclock, but find the K-Extreme Edition better to my liking. Not to show any disrespect to OCing lower lvl CPUs like i5-750s. Just like to keep my options open for ratios.
 
You must log in or register to reply here.
Back
Top