• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

So my Sabertooth came in...

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

bob4933

Member
Joined
Jan 3, 2014
"Mother of god..."



The amount of "play things" on this board is incredible. I don't even know what half this stuff does, but so far Im incredibly impressed. VRM cooling section is pretty outstanding in person comparing it to other AMD boards I see and own. Load line calibration is also outstanding. Im running 4.5ghz @ 1.35v (tested it for an hour on prime, good enough for now for me) without even a hiccup. Holds voltage very well.


Anywho, just spouting off well known features and facts, but I am quite pleased on the "step up" into a serious 990fx.
 
the saberkitty has lots of knobs and buttons and that's what makes upper end amd so much fun, get in there and mess with all of them.
 
"Mother of god..."





Anywho, just spouting off well known features and facts, but I am quite pleased on the "step up" into a serious 990fx.

Nothing like the Killer is it? Nice move bob, you'll be really happy with it. :thup:
 
Pretty much the "Go To" board for anyone that wants to OC a Fx 8/9xxx and have it work well. :thup:
 
Thanks guys. Seeing this board in action really affirms my prior statement. The killer is a gaming board. Good for mild overclock as a daily driver. Much simpler and easier to set up, but a bit less powerful as well. Probably the best matched board for the 6000 series imo for good overclocks. This sabertooth is just an entirely different echelon of power haha.


Noticed running LLC raises temps a considerable margin, is this normal?
 
More LLC means more voltage under load which means more heat produced, so yeah.
The thing to watch for on LLC is vDroop (which is an intentional effect from AMD/Intel, as a note) turning into vGain and giving you more volts than you actually asked for.
Some of the early 990fx boards had major issues with this and you'd end up with your 1.3v BIOS turning into 1.45v under load.
CPUz does a decent job of tracking CPU voltage, close enough to go with up until extreme use certainly. (Occasional massive mis-reads not withstanding. I haven't seen those in a couple generations)
 
No love for the Asrock 990fx extreme9? I'm disappointed. $20 less and still keeps pace with both top end Asus boards.

Still a good choice though Bob. Enjoy it.
 
Thanks guys. Seeing this board in action really affirms my prior statement. The killer is a gaming board. Good for mild overclock as a daily driver. Much simpler and easier to set up, but a bit less powerful as well. Probably the best matched board for the 6000 series imo for good overclocks. This sabertooth is just an entirely different echelon of power haha.


Noticed running LLC raises temps a considerable margin, is this normal?

What's the LLC doing under load? I found Ultra High to be my best setting on the CHV. It only slightly overshoots the set voltage under load and keeps it from drooping at Ultra High. It may be a bit different on the Fx 6xxx and the Sabertooth though.
 
oh, there is nothing wrong with the asrock boards, i had a fatlady, i just could not get along with the bios, for my intel rig i got an extreme9 and have just had to knuckle down and make myself get along with it.
 
Yea the BIOS is a bit of a mess. Especially with asrock wanting to be unique snowflakes and have their LLC backwards. Other than that, I love my motherboard.
 
Yea. I was informed on other motherboards, the lower the number the LLC is set to, the tighter the motherboard controls the voltage. To do the same thing on an asrock board, the higher percentage is tighter.

Setting 1 is tightest possible on say an Asus motherboard.

100% is tightest on asrock
 
Yea. I was informed on other motherboards, the lower the number the LLC is set to, the tighter the motherboard controls the voltage. To do the same thing on an asrock board, the higher percentage is tighter.

Setting 1 is tightest possible on say an Asus motherboard.

100% is tightest on asrock

I always looked at the LLC on my Asrock Z77 Extreme 4 as backwards also, compared to my 3 Asus setups.
 
Yea. I was informed on other motherboards, the lower the number the LLC is set to, the tighter the motherboard controls the voltage. To do the same thing on an asrock board, the higher percentage is tighter.

Setting 1 is tightest possible on say an Asus motherboard.

100% is tightest on asrock

I saw one user with EXT 9 some months ago maybe back first of the year and it seemed whatever bios he was using at the time had 'fixed' or changed that. I mentioned to him that the CPU_LLC was likely reversed in how we expected it to work based on having Asus boards and he went thru it and I swear it seemed not reversed any longer at least on whatever the heck bios he was on. Wish to hale I remembered who/where that was...shett shett.

I mean I know what the 'seemed/perceived' reversed CPU_LLC actions are since I have C_Ds original Fatal1ty 990FX Professional mobo now called the FatLady since I am not typing all that shett out. Hehehe. However to run that FatLady all I have to do is reverse my thinking and the Asrock FatLady does the job. Only thing I really found missing was Read To Read Delay setting in ram timings (pretty sure that is the mem timing) on my FatLady. i use that timing to make 2 mis-matched kits of G Skill ram play well together and my CHV has the needed mem timing but my FatLady does not have visible for human internevention. Otherwise the Asrock FatLady does the same job as my CHV.

I likely would not recommend in open forum to a 'prospective' buyer to go with the EXT 9 while so many of the users in here have the Sabertooth R2.0 and that is just a bigger data bank of 'what-ifs' to draw from. For myself; if needing to buy "brand-new" for an FX processor, I might go with the EXT 9 since I don't need forum help to run the numbers and can PM some of the big guns if I lose my way. Newbies and first timers are not often in that same situation.

I mention the EXT 9 as my third choice most of the time and for normally about $140.00 and for a medium overclock without the 'thought' need to adjust CPU_LLC, I would still recommend the Arsock AM3+ KilleR mobo since it works in the 4.5Ghz range without real issue it seems. Well except for the untimely death of the OPs KilleR board. But that is life.
RGone...
 
Anywho, just spouting off well known features and facts, but I am quite pleased on the "step up" into a serious 990fx.

Curious why the ASrock 990fx Fatal1ty Killer board in your sig isn't as appealing.

Have you looked into Gigabyte GA-990FXA-UD7 ?? Had one. It was a monster. Totally packed and ready for extreme cooling and overclocking.

Any How, glad you like your board. Time for the 5ghz club :welcome:
 
Curious why the ASrock 990fx Fatal1ty Killer board in your sig isn't as appealing.

Have you looked into Gigabyte GA-990FXA-UD7 ?? Had one. It was a monster. Totally packed and ready for extreme cooling and overclocking.

Any How, glad you like your board. Time for the 5ghz club :welcome:

I've been in the 5.0ghz club for a while :D

http://browser.primatelabs.com/geekbench3/367626



Its just not as stout. I can literally see the difference. If it had LLC and slightly better VRM sections like the sabertooth, it'd probably compete. It's also 50$ cheaper than the sabertooth, so each have their merits I suppose. Im actually just waiting for my new ram to come in before I push this baby any further. Hopefully that will help stability.


More LLC means more voltage under load which means more heat produced, so yeah.
The thing to watch for on LLC is vDroop (which is an intentional effect from AMD/Intel, as a note) turning into vGain and giving you more volts than you actually asked for.
Some of the early 990fx boards had major issues with this and you'd end up with your 1.3v BIOS turning into 1.45v under load.
CPUz does a decent job of tracking CPU voltage, close enough to go with up until extreme use certainly. (Occasional massive mis-reads not withstanding. I haven't seen those in a couple generations)



Its just a considerable temp difference. On my killer, 4.5 @1.4 was about 10c cooler than my sabertooth at 1.375. I'm still dinking, and yes, more voltage = more heat of course. Just kinda took me by surprise.
 
"bob4933", am just putting this here for future reference if ever needed. One brand showing X voltage and X temp versus another board showing X voltage and X temp is pretty fanciful for 'real' truth. Especially so when using software to ouput a reading.

To get these perceived differences; all the bios engineer has to do is set a slight difference in the handling of the actual voltages to the I/O chip and then the cpu monitoring software programmer must still determine what the heck they believe the voltages/temps they see, are in actuality and then present such in a manner we can read on the screen.

I will insert this here just as a semi-example of how variable is system we work with.

TDP (Thermal Design Power) and ACP (Average CPU Power)

It has been a few years now since AMD got off on another foot and began to use the
ACP (Average CPU Power) measurement with their bigger customers. AMD says that
over time they finally did gain access to the inside of one of their Opteron processors
of the FX genre as I remember it and actually got real temps internal to the processor
and not an outputted voltage. Proprietary to them alone they then came to the ACP
(Average CPU Power) rating for their big partners. With the DIY processors we still only
have a TDP spec and it is n0t by any means the full heat load that can be at the
cpu needing removal. AMD and Intel both agree that TDP is and never was given as a MAX amount of
heat energy needing to be dealt with.

The above considered we all know that we are dealing with a mirage of the heat to be
dealt with when overclocking AMD's h0t b0y FX processors. After nearly 18 months of
posts across the net from various AMD engineers, it became apparent that their core temp
was in NO way calibrated at anything less than about 40/45c under load. Not even
idling but loaded. I pay n0 attention to unloaded core temps...UNLESS I see an idling,
unloaded temp in HWMonitor of 50c. That tells me I have a cpu to cooler mis-mount of some type.

So after you wade thru all the BS and half-truths and lack of real actual information, I
would not be bothered about minimal differences from one board to the next as relates
to temps or voltages required. One slight difference in bios programming and again
we are out there in the land where the engineers are tight-lipped and un-responsive.

I have been privileged to be in from the design to release of a few motherboards. I have
run actual Alpha bioses that without the full knowledge that another board was on the
way if mine died. I will leave the name of the other person vacant but between he and I
we have had the actual release of a motherboard postponed for more bios revamping.

In many respects seeing the variances of most assembled systems using the same
parts by part number; I have wondered if a cell phone or similar was as loosely the
same as a computer system...then we would never reach the other party. Go figure.
I am sure there is something or a variable that I am interpretting wrongly, but I just
cannot comprehend the differnces from one system assembly to the next. Hehehe.
I know there are things I just know work and I go with that. All one can do overall.
RGone...ster.
 
Back