beware Manufacturer BS specs

[ik9000]

So, in finally juicing up my machine and trying to see how far I can overclock it I am finding out some interesting (and annoying) things about my motherboard.

Asus P7H55-M/USB3

My SSD benchmarks are woeful. I've noticed for a while that the SSD weren't running at the speed I expected. I thuoght it was jsut a bad bios setting. But no, having torn this thing apart it is simply mis-selling. BS. Mis-advertising, alternate facts, whatever you want to call it Asus' stated specs are ficticious:

Asus claim it has 2x sata3 (6GB/s) ports via a marvell controller. BS. They run at 250MB/s They actually run slower than the native sata2 ports!
Asus claim it has 2x PCIe 2.0 x1 but these it transpires are limited to 2.5GT/s - which means they're actually PCIE 1.0x1 ports.

How is that not false advertising? I bought the board over far more competent P55 ATX boards especially to get Sata6 and USB3. I'd have been better off with a P55 with dual GPU and a PCIe2x4 card to run USB3 from.

Having just bought a PCI card to get an extra SSD port at 500MB/s (PCI2.0x1) I now find it will be throttled to 250MB/s, in which case why bother!

I wish I had cottoned on to this sooner when trading standards might have given a damn. Sadly I can't see them doing much this many years after the event. Brand loyalty lost Asus.

[EndlessWaves]

How is that not false advertising?

It's common for hardware specifications to not require support of the highest possible speeds so that the other improvements can be implemented in situations where the higher speeds aren't possible.

You'd have to consult the specs to be certain, but they could quite easily be fully compliant implementations of SATA 6Gbps and PCI-E 2.0. Asus' specification page does say that the PCI-E slots are limited to 2.5GT/s.

[ik9000]

It's common for hardware specifications to not require support of the highest possible speeds so that the other improvements can be implemented in situations where the higher speeds aren't possible.

You'd have to consult the specs to be certain, but they could quite easily be fully compliant implementations of SATA 6Gbps and PCI-E 2.0. Asus' specification page does say that the PCI-E slots are limited to 2.5GT/s.

Saying a mobo has 2 sata 6Gbps connections that can only run at 2.5Gbps (and only at that speed one at a time) is surely false advertising. That is not sata 6Gbps, it can never attain it, under any circumstance. That is worse than SATAII speeds, which is a common complaint/observation kicking around various forums on the web.

Similarly limiting the bandwidth on a pci-e x1 and claiming it is PCIe2 is BS. It can never run at that speed. Just call it a PCI 1.0x1 and everyone knows where they stand.

Now it gets even more interesting as last night I swapped my i3 out for an i7. Suddenly a new BIOS option appears (amongst others) to choose between:

PCIE2.0x 16 and the hamstrings above OR
PCIE2.0x8 for the GPU and a boost to EITHER USB3 ports, OR to the SATA6 ports. (this follows much googling, the BIOS description is too long to read fully and the user manual is rather blank under the heading).

So again, despite not advertising any of these deficiencies the reality is that the board could only do USB3 ports if GPU was restricted below top speed, and the sata ports neutered. etc etc. None of these compromises was ever advertised, and even with IO option changed the ports still do not attain full sataIII 6Gbps. And nowhere, even in the user manual, does it hint that certain features magically appear for i7 CPUs but disappear if you use an i3.

I could have bought a much better mobo and a PCIe card on true PCIE2.0 4x and been much much better off in terms of performance.

[Dashers]

I had a similar problem on a Gigabyte X58 board. The RAID controller was run off the south bridge and there just wasn't the bandwidth to saturate disks (let alone SSDs), which made them effectively useless.

Before buying a motherboard now I download the manual and read the specs in detail for how lanes are divvied up.

[ik9000]

I had a similar problem on a Gigabyte X58 board. The RAID controller was run off the south bridge and there just wasn't the bandwidth to saturate disks (let alone SSDs), which made them effectively useless.

Before buying a motherboard now I download the manual and read the specs in detail for how lanes are divvied up.

But this is my point, the stuff about the sata6 lanes is not in the manual. Literally not there. I had to figure this out by interrogating BIOS boot screens, google and trawling through various forums circa 2010-2011 when this was a hot potato. Even the heading for IO is blank in the manual. I thought it was a typo altogether when I was running my i3 since it didn't appear at all.

Here is the Asus spec page: https://www.asus.com/Motherboards/P7...pecifications/ no clarification against Sata 6Gbps or USB3, nor the fact that PCI2x16 might be limited.
Here is the user manual: http://dlcdnet.asus.com/pub/ASUS/mb/...h55-m_usb3.pdf

Check out page 2-29. IO Level up. Highly descriptive that isn't it?

[EndlessWaves]

Similarly limiting the bandwidth on a pci-e x1 and claiming it is PCIe2 is BS. It can never run at that speed. Just call it a PCI 1.0x1 and everyone knows where they stand.

What about if you wanted to plug in a device that required the other PCI-E 2.0 improvements but doesn't use a lot of bandwidth?

Then you'd be complaining that you bought a more expensive motherboard than needed because Asus were labelling the slower PCI-E 2.0 ports as PCI-E 1.0.

Mislabelling it in a different way isn't a solution.

Of course, the ideal solution from a consumer point of view is to have all the information available, but individual marketing departments aren't keen to do that because it loses them sales.

If the Asus motherboard had said it supported PCI-E 2.0 but had a list of caveats and a similar motherboard from another brand just stated PCI-E 2.0 support then which one would you buy?

Some people would do the research and find that they both had similar levels of support in practice, but quite a few would just buy the one that hid the caveats.

They're damned if they do, damned if they don't.

The best solution is an educated customer base, if you'd realised at the time that not all PCI-E 2.0 and SATA 6Gbps implementations were running at full speed then you'd have been able to ask questions and go for the best choice instead of ending up with something you were unhappy with.

Unfortunately review sites are't always been brilliant at that. You only have to look at the number of places recommending GTX 1060s and below without mentioning the hefty G-sync price premium.

In terms of your SSD a 250MB/s cap is generally not a big deal, you might have lower sequential speeds but they don't make a great deal of difference in the real world.

[ik9000]

What about if you wanted to plug in a device that required the other PCI-E 2.0 improvements but doesn't use a lot of bandwidth?

Then you'd be complaining that you bought a more expensive motherboard than needed because Asus were labelling the slower PCI-E 2.0 ports as PCI-E 1.0.

Mislabelling it in a different way isn't a solution.

Of course, the ideal solution from a consumer point of view is to have all the information available, but individual marketing departments aren't keen to do that because it loses them sales.

If the Asus motherboard had said it supported PCI-E 2.0 but had a list of caveats and a similar motherboard from another brand just stated PCI-E 2.0 support then which one would you buy?

Some people would do the research and find that they both had similar levels of support in practice, but quite a few would just buy the one that hid the caveats.

They're damned if they do, damned if they don't.

The best solution is an educated customer base, if you'd realised at the time that not all PCI-E 2.0 and SATA 6Gbps implementations were running at full speed then you'd have been able to ask questions and go for the best choice instead of ending up with something you were unhappy with.

Unfortunately review sites are't always been brilliant at that. You only have to look at the number of places recommending GTX 1060s and below without mentioning the hefty G-sync price premium.

In terms of your SSD a 250MB/s cap is generally not a big deal, you might have lower sequential speeds but they don't make a great deal of difference in the real world.

I bought a board on the basis of it advertising 6Gbps support. I understood the need for that for SSD performance. That is educated. I checked the spec list for caveats. There were none given. I even downloaded the manual in advance. I'm not sure it is appropriate to judge what does and doesn't make a difference in the real world just because it might not affect you. I regularly have to batch save and process hundreds of high-res photographs, 102GB was the last case. That sequential read-write speed matters to me. I thought perhaps my drive had an error. Then I invested in a USB3 PCIE card thinking it might be the onboard port. Then I got a different card reader. Finally I find out after all that the problem lies in BS marketing and intentional design limitations.

I think my complaint is a valid one. You might disagree, that is up to you, but I will not be persuaded otherwise. This is false advertising so far as I am concerned, and it ought to be mandatory for all manufacturers to explain clearly any limitations or combination choices that apply to their advertised performance boasts. That would invalidate your argument above that if one misleads buyers then everyone should. We have fair advertising laws for a reason, and it is a good thing.

[Dashers]

That's a bad manual, no block diagrams. At least you can see what limitations there are on Gigabyte boards.

I think your complaint is very valid. The manual strongly suggests that you can use full speed on that chip. I'd return it as false advertising and never buy from them again. But to be fair, Asus motherboards are already fairly low down on my preferred manufacturer list.