USB 3.x PCIe Cards for Classic Mac Pro

[ActionableMango]

Any ASM1142 card should be able to do 10 Gbps

That's certainly the claim, but nobody's gotten anywhere near that on a Mac Pro that I'm aware of.

 

[h9826790]

PCIe cards with more than one controller (two or four USB controllers or a mix of USB and SATA controllers will use a PCIe switch chip. The slot negotiates a link with the switch chip. The switch chip negotiates links with the USB and SATA controllers.

If the switch chip has a 4 lane upstream port then it will eliminate the issue of x2 devices being negotiated down to x1. A PCIe switch can provide a benefit of converting fast/narrow links to equal speed slow/wide links (for example: the PCIe 2.0x2 of the ASM1142 to PCIe 1.0x4 of slot 3 or 4 of a Mac Pro 2008).

PCIe slot bifurcation is a cheaper option than a PCIe switch but bifurcation doesn't have the link conversion benefits of a PCIe switch. Also, many slots don't support bifurcation (definitely there is no Mac that supports bifurcation).

Those card "can" use switch, but not necessary "will" have its own switch onboard.

In fact, the poor performance of the card when installed in slot 2 indicated that there is no switch onboard. That's why the card unable to negotiate at PCIe 2.0 speed, but fall back to PCIe 1.0.

And this is exactly the reason why I suggested Tastannin try to install the card in slot 3 back in post #1990. His report at post #1994 agreed that my prediction is correct.

 

[joevt]

Those card "can" use switch, but not necessary "will" have its own switch onboard. In fact, the poor performance of the card when installed in slot 2 indicated that there is no switch onboard.

No. #1989 shows two Asmedia ASM2142 USB controllers each running at PCIe 3.0 speed (8 GT/s per lane) from a single slot (slot 2).

The MacPro5,1 does not have PCIe 3.0 slots. Bifurcation does not allow a PCIe 3.0 device to work at PCIe 3.0 speeds in a PCIe 2.0 slot. Therefore, a switch must be used.

Two devices on two different buses (9 and 11) cannot originate from a single slot in a MacPro5,1 using bifurcation because the MacPro5,1 doesn't support bifurcation, therefore a switch must be used.

What happens when a PCIe card using bifurcation is inserted into a slot that does not support bifurcation? I don't know. At best, only the first device can appear. At worst, no device will appear.

All pictures of the HighPoint RocketU 1344A show a PLX switch chip. A card using bifurcation would not include a switch chip. Bifurcation is used to cut cost. Adding a switch chip would negate that benefit and bifurcation has little or no other benefit.

Like I said, you need to use ioreg or IORegistryExplorer.app or pciutils to see the switch.

In fact, the poor performance of the card when installed in slot 2 indicated that there is no switch onboard. That's why the card unable to negotiate at PCIe 2.0 speed, but fall back to PCIe 1.0.

And this is exactly the reason why I suggested Tastannin try to install the card in slot 3 back in post #1990. His report at post #1994 agreed that my prediction is correct.

No evidence was given regarding the negotiated speed. There is no indication of a fall back to PCIe 1.0 speed. Poor performance was only mentioned in #1994. It was only described as "a drop in speeds when stressing both drives", but otherwise "speeds are pretty much the same as slot 2". Maybe this is a cooling problem of the card being in proximity to the graphics card?

He said there was a "noticeable speed increase" for RAID 0 in slot 3, but he didn't give any numbers so we can't compare the speeds with the expected bandwidth of the PCIe link and USB connection. The raid would have to be able to exceed 1000 MB/s to verify a proper PCIe 2.0 x4 connection. Otherwise, you need to use pciutils to see the current link speed and width of the PCIe switch and the USB controllers.

You may be correct that the card runs at PCIe 1.0 x4 in slot 2 and PCIe 2.0x4 in slot 3 because slot 1 and slot 2 (north bridge?) use a different PCIe controller than slot 3 and slot 4 (south bridge). That would be the same problem as the Amfeltec PCIe 3.0 x16 card has in my Mac Pro 2008 in slot 1 and slot 2 and we have a work around for that. I can't verify the slot 3 and slot 4 behavior because the Mac Pro 2008 only supports PCIe 1.0 in those slots. The common feature of both the Amfeltec and RocketU is a PLX PCIe 3.0 switch chip and that would make it the second half of the problem causing PCIe 1.0 fallback of the north bridge PCIe 2.0 slots. To be sure, we need the results of the setpci command to get the link speed / width for the RocketU's switch chip.

 

[jbaugh]

Sonnet Allegro Pro USB 3.0
Fresco Logic Chipset​

Today I installed the Sonnet Allegro Pro card in my cMP 5,1. The card is well made and fits tightly in its slot. This card requires no additional power connection, and driver support is built into OS X 10.8.5 or greater.

My current configuration is arranged as follows:
Slot 1 – AMD HD 7970 3GB by MVC
Slot 2 – Apple/Samsung 512GB PCIe SSD using v.1 Sintech Adaptor
Slot 3 – Sonnet Allegro Pro USB 3.0
Slot 4 – Apricorn Solo X2 w/Samsung 840 Pro SSD

The rest of my specs are in the signature at the end of this post.

After trying every USB device type that I own (Many drives (up to 4TB), webcam, USB Thumb drives, Mouse/Mice, Keyboards (4), printers, scanner, drive docking station, and USB hubs). I found no compatibility or performance issues. I transferred 16TB of data in 4TB chunks in each direction and found no issues at all. I also transferred 4TB of data in small files (again in each direction) and the process was flawless. So far I’m impressed.


About the Allegro Pro:
This card has 4 independent controllers (one for each port) and each is totally isolated from the others. The system profiler shows 4 different x1 devices installed into a single slot, it is not reported as an x4 device.

This card supports UASP and that is a specification I now demand on any USB card or storage device (except thumb drives) connected to my system. UASP is intended to speed up operations of USB connected SSDs, but I find it also brings enhanced performance to mechanical drives that are installed in UASP compatible enclosures. UASP allows you to perform multiple operations (renaming or moving other files/folders for example), while a file transfer to/from the same drive is in progress. It does this without slowing the transfer or being forced to wait for an operation to be completed. In other words, it feels like an SATA drive instead of a USB drive when manipulating files or folders.

Summary:
I my opinion, this is an exceptional card, that is slightly slower than the RocketU 1144C with similar specs. This card has additional features for charging devices such as iPhone, iPod, iPad etc… The Sonnet website recommends that you download a driver for charging support. I haven’t personally tested this, but the thread starter (ActionableMango) reports that this driver isn’t needed.

The package includes the Sonnet Allegro Pro PCIe 4-Port card & documentation. A few other details are listed below.

• 5 Year Warranty
• PCIe 2.0 x4 slot
• 4 Independent controllers
• 4 USB 3.0 Ports
• No additional power required
• No external drivers required
• UASP enabled
• 7.5w Device Charging supported on each port
• 5 Gb/s (450 MB/s) transfer rate per port
• Up to 31 devices connected at once
• OS X 10.8.5 and above supported OOTB
• Windows 7 and above supported

Conclusion:
I really like this card but it's a bit pricy at $129 from most retailers. I feel this is a great choice for the long haul with it's 5 year warranty, and device charging capabilities. I noticed that when I plug devices into this card (especially large drives), the devices are recognized and/or mounted much quicker than with the RocketU. I plan to keep this card installed permanently.

I whole heartedly recommend this card for heavy duty lifting. If your needs are less or if you are dollar wise, look at the Sonnet Allegro (non-pro version). This is a very nice addition to my cMP.

NOTE: The card pictured below is a stock photo of the v.1 card. The card that ships now has a rather large heatsink located on the large chip in the center of the board.

View attachment 565788

View attachment 565859

[doublepost=1530043693][/doublepost]My Mac Pro 2,1 maxes out with Mac OS 10.7.5. Is there any way to run this card on my machine? In other words, is there a driver that will work? Thanks.
John

 

[crjackson2134]

[doublepost=1530043693][/doublepost]My Mac Pro 2,1 maxes out with Mac OS 10.7.5. Is there any way to run this card on my machine? In other words, is there a driver that will work? Thanks.
John

This card works out of the box on every OS X / macOS since 10.8.5, however I’m not personally aware of any drivers to make it work with 10.7.5.

 

[ActionableMango]

Is there any way to run this card on my machine? In other words, is there a driver that will work?

Theoretically it should work with the generic driver in post #1. It's there in the post because somebody somewhere claims that it works.

I have not personally tried it and I make no promises. Make sure the return policy is good in case it doesn't work.

 

[h9826790]

[doublepost=1530043693][/doublepost]My Mac Pro 2,1 maxes out with Mac OS 10.7.5. Is there any way to run this card on my machine? In other words, is there a driver that will work? Thanks.
John

Upgrade to 10.8.5 or later via patch

 

[handheldgames]

Look it up on Wikipedia. USB 3.0 uses 8b/10b encoding (10 bits per byte) so 5 Gbps on the cable = 500 MB/s of data. That doesn't account for protocol overhead which will reduce that to something in the 400 MB/s range. So any benchmark that shows over 500 MB/s indicates a USB 3.1 gen 2 connection of 10 Gbps (which uses a more efficient 128b/130b encoding).





Any ASM1142 card should be able to do 10 Gbps with the correct firmware and a PCIe 2.0 slot that supports x2 (except for cards that use the ASM1142 in PCIe 3.0x1 mode which requires a PCIe 3.0 slot for full performance). Some PCIe slot controllers do not support x2 and will then force a x2 PCIe device to use x1. Examples are: the PCIe 1.0 slots of the south bridge of the MacPro3,1 Mac Pro 2008, the PCIe 2.0 slots connected directly to a Sandy bridge CPU - 2nd Generation Core Processor Family Desktop CPU. Intel's cpu and chipset documentation will show the supported PCIe lane configurations.


PCIe cards with more than one controller (two or four USB controllers or a mix of USB and SATA controllers will use a PCIe switch chip. The slot negotiates a link with the switch chip. The switch chip negotiates links with the USB and SATA controllers.

If the switch chip has a 4 lane upstream port then it will eliminate the issue of x2 devices being negotiated down to x1. A PCIe switch can provide a benefit of converting fast/narrow links to equal speed slow/wide links (for example: the PCIe 2.0x2 of the ASM1142 to PCIe 1.0x4 of slot 3 or 4 of a Mac Pro 2008).

PCIe slot bifurcation is a cheaper option than a PCIe switch but bifurcation doesn't have the link conversion benefits of a PCIe switch. Also, many slots don't support bifurcation (definitely there is no Mac that supports bifurcation).


The picture of the Highpoint RocketU 1344a shows a PLX (PEX) switch chip. The two PCIe 3.0 x2 controllers are connected to that. The switch chip will be able to translate the PCIe 3.0 x2 links of the two ASM2142 controllers to PCIe 2.0 x4 so you should be able to get full performance from each controller separately (up to 1969 MB/s), but the max from using both controllers simultaneously will be PCIe 2.0 x4 in a Mac Pro's PCIe 2.0 slot (up to 2000 MB/s).

Link speed is the speed per lane.
PCIe 4.0 = 16 GT/s (128b/130b)
PCIe 3.0 = 8 GT/s (128b/130b)
PCIe 2.0 = 5 GT/s (8b/10b)
PCIe 1.0 = 2.5 GT/s (8b/10b)

A link may use multiple lanes. This is the link width (x1, x2, x4, x8, x16, x32).

System Information.app does not show information about the bridge chip. For that you need to use ioreg or IORegistryExplorer.app to get device information, link speed and link width information. None of those will show the current actual link speed and link width. For that you need to use pciutils to examine the PCIe configuration space registers.

The switch chip is a PCIe 3.0 switch chip. In a Mac Pro, such chips might negotiate a link speed of 2.5 GT/s instead of the expected 5.0 GT/s. PCIe 1.0 x4 is still fast enough to get full performance from a single USB 3.1 gen 2 port (10 Gbps). You can't notice a problem unless you try a USB raid of 3 or 4 SSD's (> 1000 MB/s).

Read my posts in the "Amfeltec x16 PCIe with 4 SSDs: 5900+ MB/s" thread for information about PCIe 3.0 switch chips, the Mac Pro, and pciutils. #207

Not all PCIe PLX Bridge chips are created equal. In the case of the x16 8747 plx, 4x pcie 3.0 devices are mapped at 8.0 gts, not 2.5 gts.

 

[joevt]

Not all PCIe PLX Bridge chips are created equal. In the case of the x16 8747 plx, 4x pcie 3.0 devices are mapped at 8.0 gts, not 2.5 gts.

That is probably true. What uses the 48 lane 8747? 48 lanes could allow eight PCIe 3.0 x4 devices but the 8747 only has 5 ports, which allows only four PCIe 3.0 devices but at x8 width. I guess the fewer ports makes it less expensive than the 32 lane chips so someone might use it to allow four 4x PCIe 3.0 devices. It may be used on motherboards to allow four 8x slots from the CPU.

I suppose even the 32 lane 8732 of the Amfeltec could behave correctly with a new firmware update.

We don't know which PLX is used on the RocketU or if it has the same problem as the 8732 of the Amfeltec because no one has tried pciutils with it.

 

[sebo2001]

I have Mac Pro 2009 4.1 flushed to 5.1 with 2x3.4Ghz CPUs running OSX 10.10.5
I got HighPoint RocketU 1344A, card installed in PCI slot 3 system shows 8Gbps speed, mixed with OWC Mercury Elite Pro Quad RAID is super-fast, when doing disk certifications of 4 drives at the same time with SoftRAID app I can see data at 800Mbit/s which is very high for USB drive, 4 drives in Raid 0 showed 700Mbit R and 6500Mbit W speed. Problems is in RAID 5 each time I start benchmark or transfer of 100Gb of videos one of the drives disconnects… Overall very happy with the HighPoint RocketU 1344A. S.M.A.R.T data is not being reported via this card despite being USB 3.1.
It was quite expensive on Amazon.

 

[joevt]

I have Mac Pro 2009 4.1 flushed to 5.1 with 2x3.4Ghz CPUs running OSX 10.10.5
I got HighPoint RocketU 1344A, card installed in PCI slot 3 system shows 8Gbps speed, mixed with OWC Mercury Elite Pro Quad RAID is super-fast, when doing disk certifications of 4 drives at the same time with SoftRAID app I can see data at 800Mbit/s which is very high for USB drive, 4 drives in Raid 0 showed 700Mbit R and 6500Mbit W speed. Problems is in RAID 5 each time I start benchmark or transfer of 100Gb of videos one of the drives disconnects… Overall very happy with the HighPoint RocketU 1344A. S.M.A.R.T data is not being reported via this card despite being USB 3.1.
It was quite expensive on Amazon.

8Gbps = 8 GT/s? That's the link speed of a USB controller. There's two USB controllers in slot 3? What's the link width? These don't include the link speed/width of the PCIe switch chip / PCIe slot.
800 Mbit/s = 100 MB/s? Seems slow. Are these normal hard drives?
700 Mbit = 87.5 MB/s? Maybe you meant 7000 Mbit/s = 875 MB/s?
6500 Mbit = 812 MB/s? That's about the speed of two SSD's...

What chips does it use? Install pciutils and run the following commands:

sudo lspci -nn
sudo lspci -tvnn

Read my posts in the "Amfeltec x16 PCIe with 4 SSDs: 5900+ MB/s" thread for information about PCIe 3.0 switch chips, the Mac Pro, and pciutils. Link: #207
Instead of the above commands, run the pcitree.sh script (in the same page of the same post) which will also include the link speed/width of all the devices.

 

[sebo2001]

Yes 8 GT/s, and yes it was 820MB/s, in Raid 5 speed is 520MB/s

I read the post #207 but have problems running the script, I’m sure I’m doing something wrong. SSD are much faster for sure, but my main goal was to have 24TB of storage space in RAID 5.

What is interesting and frustrating when I added RocketU 1344A USB 3.1 card and upgraded to OSX 10.2.6 my eSATA enclosure FH2-SU2S2 with Newer Technology MAXPower 6G PCIe eSATA RAID Controller card stopped working. When I connect enclosure to USB all drives are mounting, but I can’t get it to work with eSATA lights are blinking but nobody is home. Disk utility does not show/detect anything.

 

[joevt]

Yes 8 GT/s, and yes it was 820MB/s, in Raid 5 speed is 520MB/s

I read the post #207 but have problems running the script, I’m sure I’m doing something wrong. SSD are much faster for sure, but my main goal was to have 24TB of storage space in RAID 5.

What is interesting and frustrating when I added RocketU 1344A USB 3.1 card and upgraded to OSX 10.2.6 my eSATA enclosure FH2-SU2S2 with Newer Technology MAXPower 6G PCIe eSATA RAID Controller card stopped working. When I connect enclosure to USB all drives are mounting, but I can’t get it to work with eSATA lights are blinking but nobody is home. Disk utility does not show/detect anything.

I think you answered two of my 9 questions but I'm not sure which of my questions the second answer is for... Send me a private message if you want help getting the scripts/commands to work. They can answer the important questions.

 

[zero hectic]

Has anyone had any issues with the Allegro Pro USB 3.0 PCIe card? I have one i my Mid 2010 Mac pro, and use it extensively with a pair of Drobo 5Ds.

Recently, the 5D stopped mounting on this card. I can plug the fine into the USB 2 ports on the mac, but no-go in the sonnet. Other drives work completely fine with the card, just not the Drobo units themselves. any suggestions?

I'm using the same card with my Mac Pro 5,1 and an older Drobo S and have never experienced an issue. Having said that, the folks at Sonnet recently pointed me to this support document from Drobo:

https://supportportal.drobo.com/retrieve/s3/knowledge/AA/AA-01339.html

**Pending further investigation, Sonnet "Allegro" PCIe USB 3.0 cards are not supported**

I'm a little concerned since I just placed an order for a Drobo 5D3 and wondering if it'll work correctly or not. I also have a Allegro USB-C PCIe card and hoping the issue doesn't apply to the card as well.

Could you give us an update? Did you manage to resolve your issue mounting your 5D?

 

[TinyWorkshop]

This card:
https://www.startech.com/eu/Cards-Adapters/USB-3.0/Cards/usb-3-1-combo-card~PEXUSB312EIC

works like a charm both with HighSierra and with Mojave.

no Bluetooth problems here

 

[joevt]

This card:
https://www.startech.com/eu/Cards-Adapters/USB-3.0/Cards/usb-3-1-combo-card~PEXUSB312EIC
works like a charm both with HighSierra and with Mojave.

Just another ASM1142 based card. Only one port of the ASM1142 is exposed (10 Gbps). The second port is used with a four port USB 3.0 hub (5 Gbps). The card is probably PCIe 2.0 x2 (electrical) since it uses a x4 connector (physical) but they are unclear about this since they only mention x1 maximums (they also mention PCIe gen 3 but I don't know if it's possible for a PCIe device to switch between PCIe 3.0 x1 and PCIe 2.0 x2 automatically?). It probably has the 10 Gbps firmware making it not useable in macOS 10.11 or earlier (except they say it has ben tested in 10.8 to 10.10 but I don't know if that's with Apple's drivers or a third party generic USB driver).

 

[TinyWorkshop]

Just another ASM1142 based card. Only one port of the ASM1142 is exposed (10 Gbps). The second port is used with a four port USB 3.0 hub (5 Gbps). The card is probably PCIe 2.0 x2 (electrical) since it uses a x4 connector (physical) but they are unclear about this since they only mention x1 maximums (they also mention PCIe gen 3 but I don't know if it's possible for a PCIe device to switch between PCIe 3.0 x1 and PCIe 2.0 x2 automatically?). It probably has the 10 Gbps firmware making it not useable in macOS 10.11 or earlier (except they say it has ben tested in 10.8 to 10.10 but I don't know if that's with Apple's drivers or a third party generic USB driver).

I don't know with macos < than 10.13

form my needs works very well and without problems

 

[weezerr]

Ive read through most of this thread.

Is there a solution to USB drives disconnecting with sleep with using the Inateck KT4004? Thanks!

Mac Pro 3,1 on 10.13.6

 

[ActionableMango]

Is there a solution to USB drives disconnecting with sleep with using the Inateck KT4004? Thanks!

Yep:

1. Ignore the disconnection notices
2. Manually eject the drives before sleep
3. Use Jettison software

Personally I chose option #1.

 

[bookemdano]

Yep:

1. Ignore the disconnection notices
2. Manually eject the drives before sleep
3. Use Jettison software

Personally I chose option #1.

Was it ever determined if cards with a molex connector for 5V power exhibited the same disconnection behavior? Is there any 5V power source inside the chassis that is maintained during sleep?

Edit: you also forgot option 4. Disable sleep entirely

That's what I did on mine since I use it at work now and they pay the electric bill

 

[ActionableMango]

Was it ever determined if cards with a molex connector for 5V power exhibited the same disconnection behavior? Is there any 5V power source inside the chassis that is maintained during sleep?

Yes, someone tested SATA power and reported "no sleep issues", just look back a couple of pages in this thread.

But I don't understand the obsession with the disconnection message. Just don't look at it. MacOS doesn't go to sleep with cached writes pending. Your data isn't lost.

Literally nobody complained until Notification Center was added to MacOS and therefore put the error messages in everyone's face.

 

[crjackson2134]

It’s never caused me a problem, and I have at least 4 connected drives all the time. I just close the notification and move on.

 

[bookemdano]

Yes, someone tested SATA power and reported "no sleep issues", just look back a couple of pages in this thread.

But I don't understand the obsession with the disconnection message. Just don't look at it. MacOS doesn't go to sleep with cached writes pending. Your data isn't lost.

Literally nobody complained until Notification Center was added to MacOS and therefore put the error messages in everyone's face.

Thanks. Missed that post.

You're on a forum for Mac fanatics and you say you don't understand the obsession over a meaningless error message?

I understand it's a low/no risk situation but at the same time if there's a way to avoid it then it's good to have that option for folks who don't want those messages to appear (or don't want to have to keep clicking them away).

Like I said before, I just avoid it by disabling sleep on my cMP, but if I was looking to buy a USB 3.0/3.1 card at this point I'd strongly consider one with a molex connector and route the SATA power from the secondary optical bay over to it.

To each his or her own.

 

[kings79]

This card:
https://www.startech.com/eu/Cards-Adapters/USB-3.0/Cards/usb-3-1-combo-card~PEXUSB312EIC

works like a charm both with HighSierra and with Mojave.

no Bluetooth problems here

About to buy this one. StarTech 5-Port USB 3.1 (PEXUSB312EIC)

For use with Pro Tools Ultimate 2018.4.1, Mac OS High Sierra & Lynx Aurora 8 Audio Interface for audio streaming I/O

Reasons;
1. Inexpensive
2. Able to connect to PCIe Aux Power for correct power for other devices
3. Has a 20 (19 used) pin connector for internal USB connection.

So;
I plan to mod the front 2x USB 2.0 Mac Pro 5,1 ports with something like this

Credit

Power can be sorted by making a cable (Not using connectors!). This is very easy and fun.


15 Pin SATA Power PinOut

6/9 Pin PCIe Power Pinout

Credit ***Make sure you test the 6pin AUX Power cable with a DMM before connecting it to the USB PCIe card.

The only thing I am missing is a source for the 15 pin (Female) SATA Connector.

What I would also do is setup the StarTech 5-Port USB 3.1 (PEXUSB312EIC) without the extra power and internal USB connection first. Then connect power, then connect internal USB.

NB Rear USB ports should be left alone because of the USB 3.0 driver booting after System boot.

Anyone have any thoughts or can point me to the Female 15 pin Connector?

 

[ActionableMango]

Anyone have any thoughts or can point me to the Female 15 pin Connector?

Get any SATA power cable and just cut off the end you don't need. As a side bonus it will already be wired. But if you really want just the connector, Mouser has them.

However, your SATA pinout shows that it expects 3.3V, 5V, and 12V. Your AUX pinout only supplies 12V. Are you going to add DC voltage converters? Or does the USB card only need 12V from that SATA connector? (That seems unlikely to me since USB is 5V.)

Back when cards required supplemental power, I pulled it in from a SATA connector in the optical bay via Y-cable. That way physical drive capacity was not reduced like it would be if I blocked a drive sled.