ASUS M4A89 Improved by v3029 BIOS

P

PcGuy78

Weaksauce
Joined
Jun 25, 2012
Messages
101
Yea, so I've been runnig w.the v2001 BIOS for ohh about 17months making due and creating success after success with each increase in MHz. Running games and transcoding data quite snapily became the norm for me once my rig was "fine tuned"....PFFFT or was it?! Um, definitely not because after flashing the CMOS chip up to v3029 a LOT of hrrrm corruption eruptions occurred w/my CFx setup indicating some incompatibility between the new "management" aka BIOS and the current components. That, coupled w.my PhenomII dropping 10C in temp overall was enough to prompt some investigating. False readings?? Nope, sensors are reporting spot on from BIOS level straight into Win7. Comparing the temps from 3 different apps and cross-referencing that data w.the mobo sensors proved the board was running much more efficiently---to say the least. Also more selectively, pointing out that while my Crossfire posted and ran 3d games above average, the BIOS I flashed on card #1 a while back actually conflicted w/card #2. Very interesting as I'd stated in the CAT12.9 BETA thread that I wasn't up for flashing my gfx becasue they seemed to be running fine. Yea right, 1st boot after flashing my MOBO and loading up ****** BETA(sry can't spill the beans xD) all I saw was incredible tearing. The 3D image was still there and actually the FPS were higher--lol--but the tearing was too much. Looked like old SEGA or NES gamed when they went haywire and blocks were everywhere but the game was intact behind the "corruption". Instead of blocks, though, it was horizontal tearing scrolling down the screen-ugly in a word. >.> Spent yesterday sifting through TPU vga database and finally found a file that seems to be working well aside from one thing. GPU #1 is @x1 supporting x16(pci-e) GPU #2 is @x16 supporting x16. Wondering if it's the chipset drivers or something which I'm looking at reinstalling. All in all the newest v3029 is pfa and shows that AMD is still hard at work instilling longevity in their products. Psh I might not even need the Vishera asap now, it's that serious. Also, remember that if you upgrade and have flashed your gfx card(s) before and they act up it's probably the "new system" weeding out any weaknesses hardware/software based. Looking forward to any comparisons. ^^v

---------------------------------------------------------------------------------------------------------------------------------------------
Quick update:
SB(AHCI,SATA RAID, etc) update and now all PCI-e lanes @x16! Ahh, a good two days of grinding software pays off!! Now for some tests!

Quick update II:
unigineSPLASH_zps3853a6fe.gif


--Score on the left is current. The three on the right represent higher CPU frequency, but were not prime stable. Wasn't satisfied w/system performance @225MHz FSB which the three on the right were using. Still fine-tuning, but from the bench we see that more gigz on the CPU help lower end FPS regardless of how subtly. However, the higher FSB also exhibits similar average fps, which makes me lean towards it for better throughput at less stress.

5770pci_e.png
xfxPCI_e.png

Both of my cards are 2.1 spec as well as my MOBO. Unsure why they're @2.0(x16 times two as opposed to x4 so not changing any drivers etc until I know what's up!) and will look into it after some time playing around w/the new setup as is--need some comparisons for later. :)

--GPU's are @875/1340 due to the gfx improving more by upping the GPU memory clock instead it's processing core. Much smoother framerates also. Bandwidth increases from 80GB/s(stock GPU memory output) up 5.8GB/s per card to 85.8GB/s per card just from that memory freq increase. Pixel&texture fillrates, conversely, increase by increasing the GPU core clock. Ill take the bandwidth over fill rates. :p

***Resolved PCI-e reading:
"Both of my cards are 2.1 spec as well as my MOBO." was incorrect. Cards are 2.1 mobo's running 2.0 PCI-e >.> Had to double check the specs online 'cuz ASUS jacked my motherboard box after an RMA. Guess everything's running up to par atm :eek: Let the tweaking begin!!?
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

M4A89 Support
 
Last edited:
The link covers GTD and TD versions of the M4A89 Series. I'm running the TD Pro USB3


--Upgraded to FX-8350: Temps and voltages are very impressive as of now. Running "Stock" 'till spring new GPU, and a decent "break-in" period. Vastly outperforms the x4 965. IMC is slightly different in that the timing it allows for 1600MHz are looser in CLK cycles but the sub-timings are tighter and DRAM driving power is increased. Also noticed my DDR3 runs at 80ohms now as opposed to 60ohms, interesting. All-in-all very pleased and using an identical platform as its predecessor provides a very stable comparison in performance and component management.
m4a89_FX.png
[/IMG]

--Another issue? Ran the latest build of p95-I think it's misinterpreting the FX CPU because as I monitored the torture w/cpu-z the bus speeds and NB--come to think of it all the frequencies were going haywire i saw 3.1ghz-5ghz on my cPU. ram went from 700mhz-900+mhz? Iono but not putting prime on it any time soon because (even tho it's for gpu still uses CPU) unigine lacked the sparaticness of prime. Using AMD-OD and a few other know suites atm to test stability etc.
 
Last edited:
I have this mobo on my AMD machine. I hate it, you have to use offset mode for the CPU voltage now with the new BIOS. Also the 16x PCI-E slot is on the bottom.
 
Board performs flawlessly and offset mode is an upgrade from standard volting because the BIOS uses factory vCore as the base reference and adding volts only requires simple addition. As for the pci-e x16 being on the bottom is true-and on the top it supports full dual-x16 lanes. A look at that sig tells me you spend less time w.this setup than I?
 
PcGuy78 said:
Board performs flawlessly and offset mode is an upgrade from standard volting because the BIOS uses factory vCore as the base reference and adding volts only requires simple addition. As for the pci-e x16 being on the bottom is true-and on the top it supports full dual-x16 lanes. A look at that sig tells me you spend less time w.this setup than I?
Click to expand...

I actually sold it not too long ago. Don't get me wrong the 1090T I had in there was running at 4GHz gave me no problems at all. It was more of an annoyance I guess. I'm glad to see that ASUS is still updating their BIOS on older motherboards though. That's one of the main reasons I always buy ASUS.

Here were my specs:
Specs:
Case - NZXT H2 Silent White
Processor - AMD Phenom X6 1090T OC'd to 3.8GHz
Motherboard - ASUS M4A89GTD Pro
RAM - Patriot Viper II DDR3 8GB
GPU - ASUS DirectCU II HD7770
HDD - Western Digital Caviar Black 1TB
PSU - Corsair HX750 Modular Power Supplu
CPU Cooler - Antec Kühler H2O 620 Water Cooler
Optical Drive - ASUS DVD-RW


I actually sold a GTX580 I had in there for the HD7770 before I sold it.
 
awww no AMD <3 in the sig then bru?? Oh, it took a long time(two years) for v3029 to release. v3027 created monstrous temp increases on the x4 965 it housed previously & I can agree that the offset presentation within that revision completely failed. Had to revert to 2001 which was installed Spring '11 and only since v3029 release has this mobo really needed an upgrade. One BIOS that I won't touch in beta stages is the MOBO's-not even attempting it; call it "build ethics". Installing beta's in the BIOS environment creates way too much room for anomolous behavior. And since they are ususally unsupported, yeah, um no thx.
 
PcGuy78 said:
The link covers GTD and TD versions of the M4A89 Series. I'm running the TD Pro USB3


--Upgraded to FX-8350: Temps and voltages are very impressive as of now. Running "Stock" 'till spring new GPU, and a decent "break-in" period. Vastly outperforms the x4 965. IMC is slightly different in that the timing it allows for 1600MHz are looser in CLK cycles but the sub-timings are tighter and DRAM driving power is increased. Also noticed my DDR3 runs at 80ohms now as opposed to 60ohms, interesting. All-in-all very pleased and using an identical platform as its predecessor provides a very stable comparison in performance and component management.
http://i170.photobucket.com/albums/u267/DMwarehouse/webvids/Pics/m4a89_FX.png

--Another issue? Ran the latest build of p95-I think it's misinterpreting the FX CPU because as I monitored the torture w/cpu-z the bus speeds and NB--come to think of it all the frequencies were going haywire i saw 3.1ghz-5ghz on my cPU. ram went from 700mhz-900+mhz? Iono but not putting prime on it any time soon because (even tho it's for gpu still uses CPU) unigine lacked the sparaticness of prime. Using AMD-OD and a few other know suites atm to test stability etc.
Click to expand...

Hi mate!

Do you have any news about Piledriver over 890FX and Prime95\OCCT\etc. stability\frequency issue?

Can you post some screens? Also some benchies would be much appreciated.

Thanks a lot for your availability.

Bye!
 
The webs got a lot of info on Piledriver architecture, not sure of what news you seek. That's the previous build of CPU-z displaying "Zambezi" instead of "Vishera" which I've updated on this end already. As far as the benching, sure, collecting various data atm and will post some comparisons to the x4 965 in the near future. As for p95 that will have to wait until I figure out why it's unstably changing frequencies. Posted about that in the above replies.
 
PcGuy78 said:
The webs got a lot of info on Piledriver architecture, not sure of what news you seek. That's the previous build of CPU-z displaying "Zambezi" instead of "Vishera" which I've updated on this end already. As far as the benching, sure, collecting various data atm and will post some comparisons to the x4 965 in the near future. As for p95 that will have to wait until I figure out why it's unstably changing frequencies. Posted about that in the above replies.
Click to expand...

Thanks for your reply. I'll stay in touch for them.
 
--Ran p95 Wed night --- Thur morn for a nice 7hr 100% load cycle. Lost a core after 4hrs(most likely the DDr3 being slightly tighter than the IMC allows because it was a rounding error and the machine was still running 7/8cores 100%). Also that sparactic behavior Prime exhibited b4 seems to have subsided for now. Strange because I'm using the same build. Loaded up CPU-z 1.58 to see what the older app diagnosed my CPU as and there were some strange readings like the TDP reporting 161w and no HT factor. Also the CPU-z 1.58 was naming the 8350 8150-P which was amusing.

--
7hr_P95.png


--
 
You should definetely try with a lower overclock and\or higher voltages. A failed Prime95, even on a single core, is a failed Prime95 :).

Thanks anyway for you effort!
 
Seven years later this CPU has crushed every speculation granted some heavy firmware/drivers were required to reach our current position--all well worth waiting for. Ill post the complete evolution soon.

Now the 7NM GPU will be the final cog in the system. Judging by the current rumors. Nah, wont repost them, but some reputable sources posted decent leaks directly from the "source". Removing SP limit really seems legit. Can we also get rid of DX soon? Just saying' because we've been bottlenecked for some time by DX-the Mantle/Vulkan benchies proved that instantly.



Also a side note on the whole Win 7 issue a.k.a RIP in about nine months support wise: [990fx/890fx rated for 600MB max/s. As we see here MUCH more is easily obtainable with proper software. And the date on the SCSI driver that simply disappeared from MS MSDN was coded in 2006 and thirteen years later we get a device that actually benefits from SCSI language in the form of NVMe. Interesting to say the least; what else are we missing from these original file systems?]

A lot of our tech is going to waste--exceedingly good--no no--great tech. I've scoured the web before posting and not many power clockers are even looking at how driving settings affect Data load/Data unload/Heat/Throughput of the NB and entire system. AMD future-proofs (almost) ALL of their CPU tech and we seem to overlook those gifts for the sake of selling something "new". I call it overproduction leading to excessive waste. That's just me though it seems. More system updates incoming

win7NVMeAMDAHCI_cacheO_2.PNG <<<--1TB WD NVMe Disk
 
Last edited:
You must log in or register to reply here.
Back
Top