Amfeltec x16 PCIe with 4 SSDs: 5900+ MB/s

[startergo]

I think the command requires sudo like this:

sudo ./fast.sh -s 00:03.0

Same thing:

-Mac-Pro:sbin g5$ sudo ./fast.sh -s 00:03.0
Password:
setpci: Unable to parse filter -s Invalid slot number.
Try `setpci --help' for more information.
# Initial PCIe 0.0 x0
setpci: Unable to parse filter -s Invalid slot number.
Try `setpci --help' for more information.
setpci: Unable to parse filter -s Invalid slot number.
Try `setpci --help' for more information.
setpci: Unable to parse filter -s Invalid slot number.
Try `setpci --help' for more information.
# Final PCIe 0.0 x0

 

[joevt]

I downloaded and looked at the script. The problem is that it thinks "-s" is the slot number. That should be removed from the command. It should look like this:

sudo ./fast.sh 00:03.0

 

[startergo]

I downloaded and looked at the script. The problem is that it thinks "-s" is the slot number. That should be removed from the command. It should look like this:

sudo ./fast.sh 00:03.0

Thanks. Now the script works.... Just doesn't do anything:
sudo ./fast.sh 00:03.0
Password:
# Initial PCIe 1.0 x16
# Final PCIe 1.0 x16

 

[joevt]

I think it has been determined that the setpci commands are not enough to alter the speed of some graphics cards.

If you're using a newer MacPro5,1 then the firmware update that comes with Mojave is supposed to allow PCIe 3.0 cards to boot at PCIe 2.0 speeds, including graphics cards, so you shouldn't need the setpci commands.

If the graphics card currently runs at PCIe 2.0 x16 (while doing 3D graphics), then the setpci commands should be able to reduce the speed to 2.5 GT/s and increase the speed back up to 5 GT/s.

 

[startergo]

I think it has been determined that the setpci commands are not enough to alter the speed of some graphics cards.

If you're using a newer MacPro5,1 then the firmware update that comes with Mojave is supposed to allow PCIe 3.0 cards to boot at PCIe 2.0 speeds, including graphics cards, so you shouldn't need the setpci commands.

If the graphics card currently runs at PCIe 2.0 x16 (while doing 3D graphics), then the setpci commands should be able to reduce the speed to 2.5 GT/s and increase the speed back up to 5 GT/s.

Well the problem is I am on the 138.0.0.0.0 firmware. The strange thing is that this card runs at 5 GT/s on the secondary slot. But also the card which is now in secondary slot runs at 5GT/s in the first slot (it is GT-120). So you see my problem:
The card appears to be OK (as it runs at 5GT/s in the secondary slot) and the slot appears to be ok (as GT-120 runs at 5GT/s in the first slot) yet my slot 1 doesn't like my Radeon or vice versa.
Edit: I have replaced the rivets of the north bridge heat sink and have discovered that now the card runs at 5GT/s in slot 1 as it is supposed to in slot 1. Not sure what changed maybe the card was not firmly plugged in the slot? Or something reset in the software due to the hardware disassembly?

 

[papadj3]

Has anyone ever heard of the Amfeltec Squid Gen 3 not working at all in a cMP 2010 (2009 Flashed to 5,1)??
The card is not recognized? I am using Innie- might that cause issues?
--- Post Merged, Yesterday at 1:30 PM ---
I am on Mojave- Should I have gotten the Highpoint??
The reason I wanted the card was that a single or double blade NOT in raid gives extra speed- I wasn't going to stripe 0??
[doublepost=1543473485][/doublepost]Amfeltec Squid Gen 3 is not recognized in slot x16 for me?!
It seems to be fine in my slot x4, tried a single SM951 AHCI boot blade alone runs at around 1400speed read/write + boots!
It works in my x4 slot but not in my x16 slot??! Please anyone having issues with Amfeltec Squid Gen 3 on cMP 2010 (2009 Flashed to 5,1) x16 slot doesn't even recognize carrier??
Canada sales says:
If Carrier board doesn’t work than it will not work in x16 as well as in x4 slot.
The problem is it looks like in your computer. This is why I recommend to check with different not so old computer. The old Mac that you have may not capable to operate with modern Gen3 board-

 

[joevt]

Has anyone ever heard of the Amfeltec Squid Gen 3 not working at all in a cMP 2010 (2009 Flashed to 5,1)??

It should work in any slot. It should be recognized even if it contains no M.2 devices - but you may need to examine ioreg or lspci to make sure it's there.

Maybe you have a problem with the x16 slot? Try the other x16 slot.

 

[papadj3]

Recognized and works in x4 slot! Not even recognized in slot x16??? I have tried another carrier in x16 slot and works fine! I use the Innie kext in Lilu- it shows external drives as internal, might that be an issue? Also, I am such a newbee, what is ioreg or lspci and how do I check it?? Other x16 slot has the GPU and works fine?? Thanks for your suggestions !

 

[startergo]

Recognized and works in x4 slot! Not even recognized in slot x16??? I have tried another carrier in x16 slot and works fine! I use the Innie kext in Lilu- it shows external drives as internal, might that be an issue? Also, I am such a newbee, what is ioreg or lspci and how do I check it?? Other x16 slot has the GPU and works fine?? Thanks for your suggestions !

the easiest way to get lspci and I/O reg is through DarwinDumper:
https://bitbucket.org/blackosx/darwindumper/downloads/
select only the necessary dumps and html report

Another lspci alternative install:
https://osxlatitude.com/forums/topic/9488-lspci-package/?tab=comments#comment-66119

 

[joevt]

Recognized and works in x4 slot! Not even recognized in slot x16??? I have tried another carrier in x16 slot and works fine! I use the Innie kext in Lilu- it shows external drives as internal, might that be an issue? Also, I am such a newbee, what is ioreg or lspci and how do I check it?? Other x16 slot has the GPU and works fine?? Thanks for your suggestions !

Did you test both x16 slots? I don't think Innie.kext is an issue.
ioreg and lspci are commands that you use in Terminal.app. ioreg gives the same information that you see using the IORegistryExplorer.app that you get with Xcode. ioreg can create a text file like this:

ioreg -l -w0 > ioreg.txt

The version of lspci that I use comes from GitHub - pciutils/pciutils: The PCI Utilitieshttps://github.com/pciutils/pciutils
Instructions are at https://forums.macrumors.com/thread...4-ssds-5900-mb-s.1936311/page-9#post-25540266
lspci can give a list of PCIe devices.

sudo lspci -nn

It can give a list of PCIe devices in a tree form:

sudo lspci -tvnn

My latest pcitree.sh script is at 2018-13-xps13-nvidia1060-6go-galax-snpr-windows-10
It is similar to the tree output of lspci but also shows the information for the bridges.

sudo ./pcitree.sh

 

[papadj3]

Thanks soo much to startergo and joevt for the time and effort you put in to start me off troubleshooting! -But- this seems way beyond my so called "technical understanding"- I will try slot #1 with squid 3 and move my GPU to slot #2-
If I also try the DarwinDumper: method if I mess up will this cause havoc for me on my computer?? I don't see the specific lspci and I/O reg to choose on the Darwin list? Will try slot#1 first- Thanks

 

[papadj3]

Tried slot 1 x16 and put GPU in slot 2x16 would not boot at all- hung at grey apple logo boot screen?
GPU must have made contact even though it sat on x4 slot next to the GPU since it is a double wide card, boot screen showed!? Again works in a x4 slot. Canada says there can't be anything wrong with card since it works in slot x4?!! Both x16 slots don't work. Obviously the card either is not recognized OR causes computer not to boot at all! Do you think the Darwin Dump will show anything that will be helpful?
[doublepost=1543515318][/doublepost]If I sell the card would be worried to buy the Highpoint- Same issues? Maybe more compatible? I don't think the flash from 4,1 to 5,1 causes issues-Do you?

 

[startergo]

Thanks soo much to startergo and joevt for the time and effort you put in to start me off troubleshooting! -But- this seems way beyond my so called "technical understanding"- I will try slot #1 with squid 3 and move my GPU to slot #2-
If I also try the DarwinDumper: method if I mess up will this cause havoc for me on my computer?? I don't see the specific lspci and I/O reg to choose on the Darwin list? Will try slot#1 first- Thanks

Select the following in the Darwin Dumper:

And then report the output of LSPCI nnvv:

 

[papadj3]

Thanks startergo- I will try Darwin and get back to you with screen shots by tomorrow!
Is Darwin Dump (Ha-I don't understand what the information means!) something I want to keep after I get the info. you need to read? Can I Dump firmware with this? Will it keep a copy of my current firmware, when trying to update new firmware? I have 140.0.0.0 now, by the way- Thanks Again! I will try this tomorrow!!!

 

[startergo]

Thanks startergo- I will try Darwin and get back to you with screen shots by tomorrow!
Is Darwin Dump (Ha-I don't understand what the information means!) something I want to keep after I get the info. you need to read? Can I Dump firmware with this? Will it keep a copy of my current firmware, when trying to update new firmware? I have 140.0.0.0 now, by the way- Thanks Again! I will try this tomorrow!!!

I was able to dump everything without the "CPU information" to the best of my knowledge (when selected that caused kernel Panic and reboot). Yes it will dump the system and video firmware as well if selected (you need to have SIP disabled for all dumps requiring root privileges)

 

[papadj3]

Select the following in the Darwin Dumper:

And then report the output of LSPCI nnvv:

Hope I got what you need! See files uploaded-
Thanks startergo !!

 

[papadj3]

Here is the last response from Canada!
"If the board is recognized in x4 PCIe slot than it has to be recognized in x16 slot.
I think you have issue with your motherboard/BIOS because it is too old."

 

[papadj3]

Thanks so much to startergo and joevt for any help you can give me regarding the lspci (nnvv).txt !!

Does anyone know what speeds I will get with 1 or 2 SM 970 blades no stripe in slot x4??

 

[startergo]

Hope I got what you need! See files uploaded-
Thanks startergo !!

from your output:

00:07.0 PCI bridge [0604]: Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port 7 [8086:340e] (rev 13) (prog-if 00 [Normal decode])
    Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
    Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR+ <PERR- INTx-
    Latency: 0, Cache Line Size: 256 bytes
    Bus: primary=00, secondary=06, subordinate=06, sec-latency=0
    Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
    BridgeCtl: Parity- SERR+ NoISA- VGA- MAbort- >Reset- FastB2B-
        PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
    Capabilities: [40] Subsystem: Intel Corporation Device [8086:0000]
    Capabilities: [60] MSI: Enable- Count=1/2 Maskable+ 64bit-
        Address: 00000020  Data: 0000
        Masking: 00000000  Pending: 00000000
    Capabilities: [90] Express (v2) Root Port (Slot+), MSI 00
        DevCap:    MaxPayload 256 bytes, PhantFunc 0, Latency L0s <64ns, L1 <1us
            ExtTag+ RBE+ FLReset-
        DevCtl:    Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
            RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop-
            MaxPayload 128 bytes, MaxReadReq 128 bytes
        DevSta:    CorrErr+ UncorrErr+ FatalErr+ UnsuppReq- AuxPwr- TransPend-
        LnkCap:    Port #7, Speed 5GT/s, Width x16, ASPM unknown, Latency L0 <512ns, L1 <4us
            ClockPM- Surprise+ LLActRep+ BwNot+
        LnkCtl:    ASPM Disabled; RCB 64 bytes Disabled- Retrain- CommClk+
            ExtSynch+ ClockPM- AutWidDis- BWInt- AutBWInt-
        LnkSta:    Speed 2.5GT/s, Width x0, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt+
        SltCap:    AttnBtn+ PwrCtrl+ MRL+ AttnInd+ PwrInd+ HotPlug- Surprise+
            Slot #128, PowerLimit 0.000W; Interlock+ NoCompl-
        SltCtl:    Enable: AttnBtn- PwrFlt- MRL- PresDet- CmdCplt- HPIrq- LinkChg-
            Control: AttnInd Off, PwrInd Off, Power- Interlock-
        SltSta:    Status: AttnBtn- PowerFlt- MRL- CmdCplt- PresDet- Interlock-
            Changed: MRL- PresDet+ LinkState+
        RootCtl: ErrCorrectable- ErrNon-Fatal- ErrFatal- PMEIntEna- CRSVisible-
        RootCap: CRSVisible-
        RootSta: PME ReqID 0000, PMEStatus- PMEPending-
        DevCap2: Completion Timeout: Range BCD, TimeoutDis+ ARIFwd+
        DevCtl2: Completion Timeout: 260ms to 900ms, TimeoutDis- ARIFwd-
        LnkCtl2: Target Link Speed: 5GT/s, EnterCompliance- SpeedDis-, Selectable De-emphasis: -6dB
             Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
             Compliance De-emphasis: -6dB
        LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete-, EqualizationPhase1-
             EqualizationPhase2-, EqualizationPhase3-, LinkEqualizationRequest-
    Capabilities: [e0] Power Management version 3
        Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
        Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-

LnkSta:    Speed 2.5GT/s, Width x0, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt+

The above is troublesome. Not sure what is going on but my best bet is to run a AHT. Or you can reflash the firmware.
@tsialex what do you think?

 

[joevt]

Thanks so much to startergo and joevt for any help you can give me regarding the lspci (nnvv).txt !!

Does anyone know what speeds I will get with 1 or 2 SM 970 blades no stripe in slot x4??

I think you should get around 1500 MB/s. You really should use an x16 slot but there's something wrong with your x16 slot.

Post results of "lspci -tvnn" with "lspci -nnvv" when you try an x4 slot.

 

[startergo]

Thanks so much to startergo and joevt for any help you can give me regarding the lspci (nnvv).txt !!

Does anyone know what speeds I will get with 1 or 2 SM 970 blades no stripe in slot x4??

I think you should get around 1500 MB/s. You really should use an x16 slot but there's something wrong with your x16 slot.

Post results of "lspci -tvnn" with "lspci -nnvv" when you try an x4 slot.

@papadj3 for this command you need to follow instructions from Joe and install pciutils

 

[papadj3]

This is way beyond me now!!

 

[startergo]

This is way beyond me now!!

When you ran the dump before it should have generated lspci tree (nnvvt).txt. Post it now. When you install the card In the x4 slot re-run DarwinDumper with LSPCI option and post both lspci tree (nnvvt).txt and lspci (nnvv).txt

 

[tsialex]

from your output:

00:07.0 PCI bridge [0604]: Intel Corporation 5520/5500/X58 I/O Hub PCI Express Root Port 7 [8086:340e] (rev 13) (prog-if 00 [Normal decode])
    Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
    Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR+ <PERR- INTx-
    Latency: 0, Cache Line Size: 256 bytes
    Bus: primary=00, secondary=06, subordinate=06, sec-latency=0
    Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ <SERR- <PERR-
    BridgeCtl: Parity- SERR+ NoISA- VGA- MAbort- >Reset- FastB2B-
        PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-
    Capabilities: [40] Subsystem: Intel Corporation Device [8086:0000]
    Capabilities: [60] MSI: Enable- Count=1/2 Maskable+ 64bit-
        Address: 00000020  Data: 0000
        Masking: 00000000  Pending: 00000000
    Capabilities: [90] Express (v2) Root Port (Slot+), MSI 00
        DevCap:    MaxPayload 256 bytes, PhantFunc 0, Latency L0s <64ns, L1 <1us
            ExtTag+ RBE+ FLReset-
        DevCtl:    Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
            RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop-
            MaxPayload 128 bytes, MaxReadReq 128 bytes
        DevSta:    CorrErr+ UncorrErr+ FatalErr+ UnsuppReq- AuxPwr- TransPend-
        LnkCap:    Port #7, Speed 5GT/s, Width x16, ASPM unknown, Latency L0 <512ns, L1 <4us
            ClockPM- Surprise+ LLActRep+ BwNot+
        LnkCtl:    ASPM Disabled; RCB 64 bytes Disabled- Retrain- CommClk+
            ExtSynch+ ClockPM- AutWidDis- BWInt- AutBWInt-
        LnkSta:    Speed 2.5GT/s, Width x0, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt+
        SltCap:    AttnBtn+ PwrCtrl+ MRL+ AttnInd+ PwrInd+ HotPlug- Surprise+
            Slot #128, PowerLimit 0.000W; Interlock+ NoCompl-
        SltCtl:    Enable: AttnBtn- PwrFlt- MRL- PresDet- CmdCplt- HPIrq- LinkChg-
            Control: AttnInd Off, PwrInd Off, Power- Interlock-
        SltSta:    Status: AttnBtn- PowerFlt- MRL- CmdCplt- PresDet- Interlock-
            Changed: MRL- PresDet+ LinkState+
        RootCtl: ErrCorrectable- ErrNon-Fatal- ErrFatal- PMEIntEna- CRSVisible-
        RootCap: CRSVisible-
        RootSta: PME ReqID 0000, PMEStatus- PMEPending-
        DevCap2: Completion Timeout: Range BCD, TimeoutDis+ ARIFwd+
        DevCtl2: Completion Timeout: 260ms to 900ms, TimeoutDis- ARIFwd-
        LnkCtl2: Target Link Speed: 5GT/s, EnterCompliance- SpeedDis-, Selectable De-emphasis: -6dB
             Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
             Compliance De-emphasis: -6dB
        LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete-, EqualizationPhase1-
             EqualizationPhase2-, EqualizationPhase3-, LinkEqualizationRequest-
    Capabilities: [e0] Power Management version 3
        Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
        Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-

LnkSta:    Speed 2.5GT/s, Width x0, TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt+

The above is troublesome. Not sure what is going on but my best bet is to run a AHT. Or you can reflash the firmware.
@tsialex what do you think?

The first thing to debug strange things is to make a BootROM dump and flash the generic MP51.fd from 10.14.1. This eliminate any hardwareID and NVRAM problems.

The problem is, a newbie user don't have the necessary knowledge on how to do this.

BTW, four different people were having problems with TITAN and TITAN X GPUs working into one Mac and not into another, three of the four were NVRAM problems that were resolved with a BootROM reconstruction, just one had backplane problems.

Another thing, early MP4,1 Mac Pros, usually those with Base_17 and Base_18 hardware descriptor, seems more susceptible for this weird compatibility problems.

 

[startergo]

This is way beyond me now!!

Can you run the DarwinDumper with the Bios-System option? This will dump your bootrom. If you are willing to follow instructions you can:
"The first thing to debug strange things is to make a BootROM dump and flash the generic MP51.fd from 10.14.1. This eliminate any hardwareID and NVRAM problems."
We can guide you