Overclock.net banner
1 - 15 of 15 Posts

· Registered
Joined
·
15 Posts
Discussion Starter · #1 ·
As the title says: all my SATA HDD/SSD are much slower on my new AM5 chipset than they used to be on my 10 years old Intel system:

  • I have a Crucial MX500 4TB which, on Intel's platform, had a sequential speed of 540MB/s but on this board it is stucked at 360MB/s (both read or write).
  • I also have two Seagate HDDs (8TB & 10TB) with a speed (on Intel) of about 200MB/s and 250MB/s respectively but on this motherboard they can't go over 150MB/s (BOTH)

This is unacceptable from my point of view and it sounds like a scam for a 400€ motherboard with a "extreme" chipset. I came accross this issue only today (after 2 months from the purchase) and I can't find a solution. Can someone please check this chipset sata speed? Has anyone with this MB had the same issue?

** I have the latest chipset drivers (AMD Chipset Driver V5.01.03.005 for Windows 10 64-bit) and BIOS 0821, which is the most stable.

** I already talked with Asus support (by phone) but they're not able to provide a concrete solution to this mess. They also gave me a very bad new about the SATA controller mounted on this board: an ancient ASM1061 (ASMedia) which is a PCIe 2.0 1x device, on a 400€ motherboard!
 

· Registered
Joined
·
15 Posts
Discussion Starter · #3 ·
Try disabling C6 sleep state and/or core parking to see if it makes a difference.
I already did it (disabling Global C-States in the BIOS). No change in SATA disk drives speed but a higher power consumption when the CPU is in idle. I guess I'm not affected by that "bug" though because my main NVME runs at full speed (7400MB/s). It's just a SATA issue.
 

· I'm NOT new to Overclock
Joined
·
12,468 Posts
Have you checked to make sure that your SATA mode is correct (ACHI over IDE or RAID)?
 

· Registered
Joined
·
383 Posts
They also gave me a very bad new about the SATA controller mounted on this board: an ancient ASM1061 (ASMedia) which is a PCIe 2.0 1x device, on a 400€ motherboard!
So let me get this straight. Instead of connecting the native SATA that the chipset supports, they instead choose to use another controller?

Font Screenshot Multimedia Brand Engineering
 

· Premium Member
Joined
·
12,082 Posts
So let me get this straight. Instead of connecting the native SATA that the chipset supports, they instead choose to use another controller?

View attachment 2605672
If that is true, the OP will only get 500MB/s (4000Mbit/s), without considering the overhead from SATA devices. Even worse, if the SATA ports share a single controller. It wouldn't surprise me if the vendor skimped on such "legacy" interfaces like SATA ports. Still, SATA performance on AMD chipsets has always been a bit lackluster compared to competing Intel chipsets.
 

· Registered
Joined
·
15 Posts
Discussion Starter · #8 ·
So let me get this straight. Instead of connecting the native SATA that the chipset supports, they instead choose to use another controller?
I'm afraid it is but I can't figure out the reason of such a crazy decision on a Rog Strix motherboard. It's like if there was a sort of cap on the maximum speed the SSD can reach (400MB/s) as well as the HDD (160MB/s). Someone on youtube should consider to test this motherboard focusing on this issue because Asus can't get away with it.
 

· Overclocking Enthusiast
Joined
·
7,080 Posts
Removing the sata ports opens up pcie 3.0 lanes. However i don't really see anything that stands out. That might need those extra lanes.

The ASM1061 is a 2 port controller. It's not uncommon to see it used to add 2 additional sata ports. Are all of the ports ASM1061?
 
  • Rep+
Reactions: kokoslon5

· Iconoclast
Joined
·
34,371 Posts
They also gave me a very bad new about the SATA controller mounted on this board: an ancient ASM1061 (ASMedia) which is a PCIe 2.0 1x device, on a 400€ motherboard!
It looks like it uses two ASM1061s for SATA PHYs, with two ports connected to each:


Best performance with your combination of SATA devices is probably with the HDDs sharing one pair of ports and the SSD alone on the other pair.

Anyway, SATA performance was clearly an afterthought on this board.

Removing the sata ports opens up pcie 3.0 lanes. However i don't really see anything that stands out. That might need those extra lanes.

The ASM1061 is a 2 port controller. It's not uncommon to see it used to add 2 additional sata ports. Are all of the ports ASM1061?
They all look like they are on ASM1061s.

At first I thought they had maybe routed the SATA lines to every M.2 slot, but the board makes no mention of SATA support on the M.2s.

I don't know if they needed two extra lanes for something else, or if they simply chose to only route two lanes from the chipset to save area on an already cluttered board without having to use more PCB layers.

It's also possible that these older controllers are more tolerant of PCI-E overclocking and the Strix E's don't seem to have an external clock generator, so this could be important for BCLK OCing.
 

· Registered
Joined
·
15 Posts
Discussion Starter · #11 ·
Best performance with your combination of SATA devices is probably with the HDDs sharing one pair of ports and the SSD alone on the other pair.
I tried this way but I saw no changes in speeds. I tried also to transfer big files (40-50GB each) from the Crucial MX500 to the Samsung 990Pro (which is on the PCIe 5.0 M2 slot) but I got the same cap. Just for info: do you think the Rog Strix X670E-E (which is almost the same board) has this issue as well?

They all look like they are on ASM1061s.
I don't know if they needed two extra lanes for something else, or if they simply chose to only route two lanes from the chipset to save area on an already cluttered board without having to use more PCB layers.
Yes, all of them is on that crappy controller and the "funny" fact is that the site where you got that picture (laurent's choice) titled the review "The only STRIX worth owning!" o_O
 

· Registered
Joined
·
383 Posts
I tried this way but I saw no changes in speeds. I tried also to transfer big files (40-50GB each) from the Crucial MX500 to the Samsung 990Pro (which is on the PCIe 5.0 M2 slot) but I got the same cap. Just for info: do you think the Rog Strix X670E-E (which is almost the same board) has this issue as well?
It should not but who knows. If we look into the manual of the B650 board it says

Font Circle Rectangle Number Pattern


That seems to be the 4x PCIe lanes from the pic I posted earlier. If we look into the X670 manual it says just the same:

Mammal Font Circle Rectangle Number



So I would hope they choose the 4x PCIe + 4x SATA option as described in the AMD diagram.

Font Technology Screenshot Rectangle Brand
 

· AMD OC'ing Enthusiast
Joined
·
3,585 Posts
OP jumps immediately to blaming the motherboard. Excitement of the new system took over. Lets check a few things out first!!


Was the old Intel rig the EXACT same operating system? Minus obvious driver differences? If ported over directly, clean all Intel drivers out now, or do a fresh install.

Any type of backups being done, or imaging processes taking place that might have accidentally been set with different optimizations?

Is the BIOS reading the drives correctly? On my old X99 Asrock board, you could actually tell the storage controllers what type of device (ssd or hdd) they were connected to, to force the correct behavior.


I am of the opinion that a deep dive into all the drivers will be necessary.


As far as I am concerned, the AMD AHCI drivers have always been beneath/lesser of the default windows drivers in terms of performance. The intel machine likely had the Intel optimized drivers, which do enable BETTER performance.

That said, synthetic benchmarks do not always tell the correct tale. Ive dealt with ssd's that benched higher than others, but had recorded game loading that showed the opposite.



Some hardware just doesnt like co-operating properly!


Best of luck. Computers can be frustrating at times.
 

· Iconoclast
Joined
·
34,371 Posts
Just for info: do you think the Rog Strix X670E-E (which is almost the same board) has this issue as well?
No, the 670E-E appears to have the SATA ports connected directly to one of the Promontory chips (nothing big enough in the vicinity of the ports to be a 3rd party controller):

EDIT -- swapped out old image for a better picture.

Yes, all of them is on that crappy controller and the "funny" fact is that the site where you got that picture (laurent's choice) titled the review "The only STRIX worth owning!" o_O
SATA performance is a complete non-issue for most of the target market. They could probably get away with providing only a single port.

OP jumps immediately to blaming the motherboard.
The board's SATA solution is laid out in such a way that will harm sequential transfers on almost any SSD as well as faster HDDs that are attached to the same PHY.

I wouldn't personally consider it a deal breaker, but the performance observed is almost certainly the board's fault. Still, always wise to rule out other issues, especially if SATA performance is critical enough for the OP to want to swap boards over it.
 
  • Rep+
Reactions: Sir.TEO

· Registered
Joined
·
15 Posts
Discussion Starter · #15 ·
OP jumps immediately to blaming the motherboard. Excitement of the new system took over. Lets check a few things out first!!
I'm not blaming my whole motherboard but it's a matter of fact that this SATA-issue is a clear and big fault for a top of the range B650E motherboard which costs a lot of money.

Was the old Intel rig the EXACT same operating system? Minus obvious driver differences? If ported over directly, clean all Intel drivers out now, or do a fresh install.
Yes, same OS (Windows 10 64bit). It's not a ported OS but a fresh install and I assembled this system two months ago.

Any type of backups being done, or imaging processes taking place that might have accidentally been set with different optimizations?
No, nothing at all.

Is the BIOS reading the drives correctly? On my old X99 Asrock board, you could actually tell the storage controllers what type of device (ssd or hdd) they were connected to, to force the correct behavior.
I already forced all the devices to the correct type

That said, synthetic benchmarks do not always tell the correct tale.
As I said in my previous post I haven't used only synthetic benchmark but I did direct tests.

Computers can be frustrating at times.
This ZEN4 rig certainly is but everything is working properly after my tune, everything but SATAs

__

@wiciaki I'm looking forward to seeing your tests
 
1 - 15 of 15 Posts
Top