DDR3 Skylake board?

[Guest]

So far in my build I've been focusing on cheap H110 mATX Skylake motherboards, but I found a B150 ATX board for not much more. I'd prefer the b150 for the extra PCIE and other connectors, but it only supports DDR3. Will that be an issue, as it does limit my upgrade path to one generation rather than 2, as only Skylake works with DDR3/4 and kabylake only works with DDR4?

[Davidwr80]

As far as I know, Kaby lake does support DDR3 too.

[2FA]

It technically only supports DDR3L which is the low voltage variant of DDR3 and limits your options when purchasing a kit. I would just get a DDR4 even if it costs a little bit more.

[Davidwr80]

1 minute ago, DeadEyePsycho said:

It technically only supports DDR3L which is the low voltage variant of DDR3 and limits your options when purchasing a kit. I would just get a DDR4 even if it costs a little bit more.

there are 1151 boards that support full DDR3 though, not only DDR3L

[2FA]

1 minute ago, Davidwr80 said:

there are 1151 boards that support full DDR3 though, not only DDR3L

The OP didn't say the model so we don't know if it does or not.

[Guest]

6 minutes ago, DeadEyePsycho said:

The OP didn't say the model so we don't know if it does or not.

Sorry, it's the GA-B150-HD3 DDR3. 

[Guest]

13 minutes ago, DeadEyePsycho said:

The OP didn't say the model so we don't know if it does or not.

But for some reason one site says it has 2 PCIE x16, the other I checked says it has 1 and a PCI slot? The gigabyte website says it has 6 audio connectors as well as 2 PCIE x16 so maybe I'm looking at different motherboards her but idk?

[MageTank]

5 minutes ago, Davidwr80 said:

there are 1151 boards that support full DDR3 though, not only DDR3L

 

3 minutes ago, DeadEyePsycho said:

The OP didn't say the model so we don't know if it does or not.

Any board that supports DDR3L supports DDR3. There is no board that I am aware of that refuses to post with a DDR3 kit if it's capable of using DDR3L. These are the list of DDR3 1151 boards: http://pcpartpicker.com/products/motherboard/#s=30&L=3&sort=a8&page=1

 

You will not find a single board on this list that cannot support 1.5v DDR3. To address the giant elephant in the room: Yes, Intel said going above 1.35v will damage the IMC. You know what else Intel said? Going above 1.5v on Haswell would damage it's IMC. Don't believe me? Here's their own whitesheets:

Skylake: http://www.intel.com/content/www/us/en/processors/core/desktop-6th-gen-core-family-datasheet-vol-1.html (Go to page 116)

Haswell: http://www.intel.com/content/www/us/en/processors/core/4th-gen-core-family-desktop-vol-1-datasheet.html (Go to page 103)

 

You will see this for Skylake:

Notice how it says Max VDDQ (VDIMM) = 1.35 +5%? This means your max "safe" VDIMM is 1.4175v. Pretty scary, huh? Now let's look at Haswell:

Max is 1.5 + 5%. This means 1.575v is the max "save" VDIMM. Now tell me, how many people do you know of that ran 1.65v kits on their Haswell CPU's for years, without any issues? Under load, you typically see spikes of 1.7v on VDIMM on those kits. A massive 13.4% increase over their "safe" number. Meanwhile, 1.5v is only 11.2% over Skylakes "safe" number. To further put this myth to rest, I've been running 1.65v memory on a Core i5 6600T for about 8 months now, and nothing has happened. No degradation, no random crashes, nothing. Chip still overclocks fine at 4.59ghz, and the memory itself is a super tight 2133 C9 kit with tight tertiaries. If anything, that IMC is under heavier stress than any normal XMP profile could ever provide. 

 

My point is, Intel's standards are silly. We've all broken them in the past, and nothing bad happened then. Why everyone makes a big deal out of it, simply because it's Skylake, is beyond me. Besides, I run my 3600 C14 DDR4 kit at 1.4v, which spikes to 1.44v under load. This is already outside of Intel's safe specs, and it's also doing just fine. If VDIMM could kill a CPU, i'd be the first to know. 

[Guest]

5 minutes ago, MageTank said:

 

Any board that supports DDR3L supports DDR3. There is no board that I am aware of that refuses to post with a DDR3 kit if it's capable of using DDR3L. These are the list of DDR3 1151 boards: http://pcpartpicker.com/products/motherboard/#s=30&L=3&sort=a8&page=1

 

You will not find a single board on this list that cannot support 1.5v DDR3. To address the giant elephant in the room: Yes, Intel said going above 1.35v will damage the IMC. You know what else Intel said? Going above 1.5v on Haswell would damage it's IMC. Don't believe me? Here's their own whitesheets:

Skylake: http://www.intel.com/content/www/us/en/processors/core/desktop-6th-gen-core-family-datasheet-vol-1.html (Go to page 116)

Haswell: http://www.intel.com/content/www/us/en/processors/core/4th-gen-core-family-desktop-vol-1-datasheet.html (Go to page 103)

 

You will see this for Skylake:

Notice how it says Max VDDQ (VDIMM) = 1.35 +5%? This means your max "safe" VDIMM is 1.4175v. Pretty scary, huh? Now let's look at Haswell:

Max is 1.5 + 5%. This means 1.575v is the max "save" VDIMM. Now tell me, how many people do you know of that ran 1.65v kits on their Haswell CPU's for years, without any issues? Under load, you typically see spikes of 1.7v on VDIMM on those kits. A massive 13.4% increase over their "safe" number. Meanwhile, 1.5v is only 11.2% over Skylakes "safe" number. To further put this myth to rest, I've been running 1.65v memory on a Core i5 6600T for about 8 months now, and nothing has happened. No degradation, no random crashes, nothing. Chip still overclocks fine at 4.59ghz, and the memory itself is a super tight 2133 C9 kit with tight tertiaries. If anything, that IMC is under heavier stress than any normal XMP profile could ever provide. 

 

My point is, Intel's standards are silly. We've all broken them in the past, and nothing bad happened then. Why everyone makes a big deal out of it, simply because it's Skylake, is beyond me. Besides, I run my 3600 C14 DDR4 kit at 1.4v, which spikes to 1.44v under load. This is already outside of Intel's safe specs, and it's also doing just fine. If VDIMM could kill a CPU, i'd be the first to know. 

OK thanks