Titan X Pascal - Abnormal boost clock behavior?

Echelon-

n00b
Joined
Mar 1, 2017
Messages
8
Hey Everyone,

To my understanding the base clock of the 1080ti will be 1480MHz and the boost clock will be 1582MHz, currently I'm getting stock on my Titan XP a base clock of 1418, but a boost anywhere from 1,797 MHz-1,886 MHz on air and even 1900mhz+ or so on water without touching any settings.

Is this normal for the dynamic nature of the boost clock and ambient temps / TDP / power allowing things to go higher than normal?

Long version:

A couple weeks ago I sold my old Titan X Maxwell and got a Titan X Pascal in exchange.
I've got 2 weeks left of a return period for the Titan XP. Outside of saving some cash, I am debating if it's worth it performance wise to return the Titan XP and get a Founder's Edition 1080ti.

Oddly, my Titan XP on air and stock settings before even having afterburner installed would report as a 1418 base core speed, but when boost kicked in with gaming load or 3DMark, it'd go anywhere between 1,797 MHz-1,886 MHz.
(3DMark and Afterburner would report the same high boost clocks, I thought it was a misreading at first)

It was to my understanding the Titan XP Pascal should boost the core clock from a base of 1418Mhz up to 1531Mhz stock depending if ambient temperature / TDP etc allowed and no downclocking occurred but I'm seeing different stock results.

Overclocked, at 100% Power Limit with watercooling with no other changes other than adjusting clockrates, I'm able to achieve a +200mhz memory clock increase and +200mhz Core clock which results in 2050MHz-2126 MHz stable.

Is this expected normal operation, or did I get lucky with already running at stock higher than a 1080ti when boost kicks in?

Thanks
 
I turn my fan up as I don't have delicate little ears :) , and get the same , easily 2000mhz with no crashes and haven't tried to push it more but looks like I could. I've bumped up the temp and power a little but haven't even needed to as monitors show it still never approaches the default throtttle temp as is.

Glad they are doing the 1080ti for folks , but still quite happy with my Titan XP and I really don't buy that the Ti will be much faster at all.
 
I'm also curious about the 1080ti's power handling, with the re-engineered power design that has a 7-phase dual field-effect transistor (FET) and 14 dual-FETs architecture as stated by them. I know it lowers temps by 5c or so from the original (which currently are not a problem for me on water at all). But would that allow for higher OCs and better power delivery over the TitanXP?
I also saw they tout 11Gbps G5X memory, where I think the TitanXP is at 10? But I think that's to make up for the lower bit memory interface unless I'm wrong?

Just weighing my options a bit, It should be interesting to see the benchmark results incoming soon.
 
If it's an easy return and swap it'd be worth it I think for the 500ish savings , I genuinely think that for all practical purposes it will be a wash between the two performance wise once results are out with OC'd Titan XP vs OC Ti.

I guess the only factor is if you could get one on first round of orders or if you don't mind being without a card if not for a couple weeks.
 
Hey Everyone,

To my understanding the base clock of the 1080ti will be 1480MHz and the boost clock will be 1582MHz, currently I'm getting stock on my Titan XP a base clock of 1418, but a boost anywhere from 1,797 MHz-1,886 MHz on air and even 1900mhz+ or so on water without touching any settings.

Is this normal for the dynamic nature of the boost clock and ambient temps / TDP / power allowing things to go higher than normal?

Long version:

A couple weeks ago I sold my old Titan X Maxwell and got a Titan X Pascal in exchange.
I've got 2 weeks left of a return period for the Titan XP. Outside of saving some cash, I am debating if it's worth it performance wise to return the Titan XP and get a Founder's Edition 1080ti.

Oddly, my Titan XP on air and stock settings before even having afterburner installed would report as a 1418 base core speed, but when boost kicked in with gaming load or 3DMark, it'd go anywhere between 1,797 MHz-1,886 MHz.
(3DMark and Afterburner would report the same high boost clocks, I thought it was a misreading at first)

It was to my understanding the Titan XP Pascal should boost the core clock from a base of 1418Mhz up to 1531Mhz stock depending if ambient temperature / TDP etc allowed and no downclocking occurred but I'm seeing different stock results.

Overclocked, at 100% Power Limit with watercooling with no other changes other than adjusting clockrates, I'm able to achieve a +200mhz memory clock increase and +200mhz Core clock which results in 2050MHz-2126 MHz stable.

Is this expected normal operation, or did I get lucky with already running at stock higher than a 1080ti when boost kicks in?

Thanks


I don't know about water, but on my titan p pretty much the same as yours with air.

Boost clocks are dynamic based on ambient temps and air flow of your case so yeah they can vary a lot.
 
Dahkoht, at stock on air with no adjustments, did your boost perform the same? Getting around 1800 or so? (With no core increase or power %) You mentioned pushing things, so that led me to believe you OC'd

Haha, being without a card is a no go! I bet they'll sell out quick.
Yeh, going back to monetary reasons it is worth saving 500ish, but I did make my money back on my Titan X Maxwell luckily, so it'd be like getting extra money! I'm thinking out side of that for now though. Pure performance is the only equation at the moment as it's pretty curious.
 
Dahkoht, at stock on air with no adjustments, did your boost perform the same? Getting around 1800 or so? (With no core increase or power %) You mentioned pushing things, so that led me to believe you OC'd

Haha, being without a card is a no go! I bet they'll sell out quick.
Yeh, going back to monetary reasons it is worth saving 500ish, but I did make my money back on my Titan X Maxwell luckily, so it'd be like getting extra money! I'm thinking out side of that for now though. Pure performance is the only equation at the moment as it's pretty curious.


Voltage is automatically increased with Boost 3.0.
 
Yes, this is normal. Mine boosts to around 1850 MHz with the blower cooler and without touching anything else. I can overclock it to about 2050 MHz and it will settle to around 1980 MHz during a gaming session. It's not unusual to see a 2.1 GHz overclock on a Titan X cooled with water.
 
Ah, this is all pretty interesting. I had no idea about Boost 3.0 adjusting Voltage.

The whole "Faster than a Titan X" thing doesn't make much sense with all this new info to me, performance wise at stock it seems like it'll be pretty much in line with Titan XP performance on average, just of course at a much lower cost.
 
Ah, this is all pretty interesting. I had no idea about Boost 3.0 adjusting Voltage.

The whole "Faster than a Titan X" thing doesn't make much sense with all this new info to me, performance wise at stock it seems like it'll be pretty much in line with Titan XP performance on average, just of course at a much lower cost.
Well notice they didn't say how much faster. In my opinion reviews will probably show it is only faster because of the higher base clock. Will be interesting to see if there will be any difference at all when both are overclocked to 2.0-2.1 GHz.
 
Ah, this is all pretty interesting. I had no idea about Boost 3.0 adjusting Voltage.

The whole "Faster than a Titan X" thing doesn't make much sense with all this new info to me, performance wise at stock it seems like it'll be pretty much in line with Titan XP performance on average, just of course at a much lower cost.

Maybe they're winding up for a Titan X Pascal 2?
 
Well notice they didn't say how much faster. In my opinion reviews will probably show it is only faster because of the higher base clock. Will be interesting to see if there will be any difference at all when both are overclocked to 2.0-2.1 GHz.


They should be near identical, the only difference is in theory Titan X should have more memory bandwidth when the RAM is overclocked because of the slightly wider bus, but I don't know how many games are that sensitive to memory bandwidth when it's already in the 400GB/s range.
 
Ah, this is all pretty interesting. I had no idea about Boost 3.0 adjusting Voltage.

The whole "Faster than a Titan X" thing doesn't make much sense with all this new info to me, performance wise at stock it seems like it'll be pretty much in line with Titan XP performance on average, just of course at a much lower cost.

I reckon it will be slightly faster at stock speeds but wont have as high % overclock so they will level out at max.
 
Back
Top