*Official Benchmark Thread* - Post it here or it didn't happen :D

[Mr. Fox]

2 minutes ago, johnksss said:

I'm still tripping out on the voltage. it's 1.110V! So a shunt mod would actually do this card justice or an unlocked power mod PNY vbios. I had not seen any other card with a voltage that high before.

That's excellent. If it's 1.100V without even using the 100mV slider that is very good news, especially if using it takes it up to 1.200V. The 4090 Suprim and Gigabyte 4090 Gaming OC are both 1.100V max with the extra 100mV, which isn't enough.

[Reciever]

11 minutes ago, Mr. Fox said:

I don't mind a game having options available to use a controller. I know some people prefer that and I'm okay with whatever floats their boat. Gaming should be done in whatever matter the gamer wants it. What I don't like is having no controller connected and being forced to work my way through configuration menus that assume I have a controller attached even though I don't, and all of the on screen action prompts show me controller buttons to press even though no controller is attached. There are a lot of console games ported to PC that don't have that kind of a shortcoming.

---

That's awesome brother @johnksss I'm super glad to hear that you're pleased. 

 

PNY has my attention and I expect to see great things develop with their brand with Vince involved. If given the chance I'm pretty sure he's going to take them to greatness like he did with EVGA 

 

 

IIRC there are options for that in the settings.

[johnksss]

8 minutes ago, Mr. Fox said:

That's excellent. If it's 1.100V without even using the 100mV slider that is very good news, especially if using it takes it up to 1.200V. The 4090 Suprim and Gigabyte 4090 Gaming OC are both 1.100V max with the extra 100mV, which isn't enough.

Without the slider, its 1.085V and with slider, it's 1.110V

And my PCIE slot voltage and 16-Pin voltage go no lower than 12.1. While my other card dipped in the 11.8 to 11.9 range.

Please be a good card.😂

[tps3443]

1 hour ago, johnksss said:

I'm still tripping out on the voltage. it's 1.110V! So a shunt mod would actually do this card justice or an unlocked power mod PNY vbios. I had not seen any other card with a voltage that high before.

That’s really good! Which card is this?
 

I would absolutely shunt the card. I went with the 1.5 Ohm 3 watt resistor for my FE from mouser. The stock FE has a single 1 ohm resistor, so stacking 1.5 ohm on top will give me 1,050 watts not counting the PCIe slot power. Bitspower may or may not be sending me a GPU block for testing. Fingers crossed.🤞 I’ve been bugging them to death so we’ll see what happens. 
 

The 5090 even shunted is going to chow down 900+ watts even with lower voltages. These cards are throttling hard though man. The biggest issue is the internal clock that is actually running. The cards are reflecting a higher clock speed externally, but it’s really running much slower than that effectively. It’s like a rubber band effect to prevent crashing due to power limits. So shunting will eliminate this entirely! 
 

 

[johnksss]

6 minutes ago, tps3443 said:

That’s really good! Which card is this?
 

I would absolutely shunt the card. I went with the 1.5 Ohm 3 watt resistor for my FE from mouser. The stock FE has a single 1 ohm resistor, so stacking 1.5 ohm on top will give me 1,050 watts not counting the PCIe slot power. Bitspower may or may not be sending me a GPU block for testing. Fingers crossed.🤞 I’ve been bugging them to death so we’ll see what happens. 
 

The 5090 even shunted is going to chow down 900+ watts even with lower voltages. These cards are throttling hard though man. The biggest issue is the internal clock that is actually running. The cards are reflecting a higher clock speed externally, but it’s really running much slower than that effectively. It’s like a rubber band effect to prevent crashing due to power limits. So shunting will eliminate this entirely! 
 

 

Before I can get to shutting it it has to be a good card to work with. And right now it's hard to say as at stock it's pulling 630W

[tps3443]

4 minutes ago, johnksss said:

Before I can get to shutting it it has to be a good card to work with. And right now it's hard to say as at stock it's pulling 630W

Well, I’d look at undervolting to give you an idea. It could be pulling more just due to a higher default voltage curve and that’s making it seem like it’s using much more power. Is this a PNY RGB OC model? What was wrong with the Trio you got? 

[johnksss]

9 hours ago, tps3443 said:

Well, I’d look at undervolting to give you an idea. It could be pulling more just due to a higher default voltage curve and that’s making it seem like it’s using much more power. Is this a PNY RGB OC model? What was wrong with the Trio you got? 

That's what im trying to do right now, but voltage seems to be all over the place...

 

Edit:

It's the the Epic RGB OC model. And the Gaming trio someone who could not get a card bought it off me.

[tps3443]

I can’t believe we have an Oblivion remastered in 2025 on PC it has RT/FG/DLSS support lol! Just came out today on game pass. I remember struggling to run this one smoothly at 20-30fps with my GeForce 6600GT. I was running 1280x1024. 
 

I ended up beating Oblivion on a ATI X850 Pro that was flashed to an X850XT bios via floppy disc lol.

[electrosoft]

3 hours ago, johnksss said:

Without the slider, its 1.085V and with slider, it's 1.110V

And my PCIE slot voltage and 16-Pin voltage go no lower than 12.1. While my other card dipped in the 11.8 to 11.9 range.

Please be a good card.😂

 

Hoping it's a banger @johnksss, but then again with your past work on so-so/average hardware.....

 

 

2 hours ago, tps3443 said:

That’s really good! Which card is this?
 

I would absolutely shunt the card. I went with the 1.5 Ohm 3 watt resistor for my FE from mouser. The stock FE has a single 1 ohm resistor, so stacking 1.5 ohm on top will give me 1,050 watts not counting the PCIe slot power. Bitspower may or may not be sending me a GPU block for testing. Fingers crossed.🤞 I’ve been bugging them to death so we’ll see what happens. 
 

The 5090 even shunted is going to chow down 900+ watts even with lower voltages. These cards are throttling hard though man. The biggest issue is the internal clock that is actually running. The cards are reflecting a higher clock speed externally, but it’s really running much slower than that effectively. It’s like a rubber band effect to prevent crashing due to power limits. So shunting will eliminate this entirely! 
 

 

 

I knew your Founders stood no chance of staying in its pristine condition! 🤣

 

Like I said many days ago when you were just talking about adding pads or something along those lines....

 

"and it starts." 😁

 

Definitely can't wait to see it shunt modded and blocked bro!

 

 

[jaybee83]

12 hours ago, tps3443 said:

I haven’t seen any issues with the latest drivers. I use my PC for 8 hours a day a least. I’m not sure what the deal is. Latest driver seems fine, it reduced idle clocks even further down to 22Mhz from 180Mhz. I’m not really checking any fps differences in games, maybe I should I suppose. I saw a slight improvement in Steel Nomad. Card can still reach max 1075mv. Temps sensor seems to work okay as well. I have never experienced black screens, if I add like +475 you will get a black screen from a crash. But other than that the drivers seem perfectly fine.

 

https://www.3dmark.com/sn/5548902

 

welp thats the sucky thing, it can get you out of nowhere, even if u had no issues beforehand whatsoever. im happily chugging along on 572.83 while others are crash n burning on it 😄 shouldnt be this way...this is not the way

 

11 hours ago, tps3443 said:

My last driver 572.83 definitely needed a hard reset to recover from. 

 

really sounds more and more like its driver-related then. was the hard reset overclocking related?

 

11 hours ago, johnksss said:

Thank you @Mr. Fox ! Package arrived safe and sound and it up and running like a champ!

 

Okay, so i ended up selling my Gaming Trio 5090 to one of my customers and grabbed this PNY Epic from Microcenter a little while ago. And I must say i was quite surprised at it's stock bios settings.

 

 

11 hours ago, johnksss said:

I'm still tripping out on the voltage. it's 1.110V! So a shunt mod would actually do this card justice or an unlocked power mod PNY vbios. I had not seen any other card with a voltage that high before.

 

9 hours ago, johnksss said:

Before I can get to shutting it it has to be a good card to work with. And right now it's hard to say as at stock it's pulling 630W

 

9 hours ago, johnksss said:

That's what im trying to do right now, but voltage seems to be all over the place...

 

duuuuude, that card looks crazy! mine only does 1.01V at stock with clocks around 2825 Mhz, with voltage slider maxed it goes up to 1.025V 😄 those voltages are insane! ive only seen maybe 1-2 other cards that can break through 1.1V, u definitely have a special one there. 3000+ Mhz at stock, crazy...

 

and yes ive noticed the same, voltages are all over the place, heavily depending on what ure running and even fluctuate wildly during sessions / bench runs. thats even the case if the power limit hasnt even been reached yet. not sure whats going on, hope this is smth that can be rectified with drivers / vbios updates....

 

overall, its total chaos when it comes to clocks and voltages, every card is so different among 5090s, both in wattage/clocks/voltage, way more variance than with the 4090.

 

those 630W, are those transient spikes or peak load? i might have to give your vbios a go, pls share with the class! 😄 

 

keep us in the loop on your findings, highly interested 🙂 

[tps3443]

1 hour ago, electrosoft said:

 

Hoping it's a banger @johnksss, but then again with your past work on so-so/average hardware.....

 

 

 

I knew your Founders stood no chance of staying in its pristine condition! 🤣

 

Like I said many days ago when you were just talking about adding pads or something along those lines....

 

"and it starts." 😁

 

Definitely can't wait to see it shunt modded and blocked bro!

 

 

I really wanted my whole system to be watercooled when I was building this rig and I had kinda prepped with the 4090. We just don’t have any blocks yet for FE models. So, I had already ordered my Fujipoly fullcover rear thermal pad (Originally for the 4090 never used it) but now I’ve ordered some shunts, but I can’t add any of these items until I get a waterblock. FE blocks are taking a little longer than the other models, but they’re coming. I’m really curious what the Bitspower block will even look like. 

[electrosoft]

4 hours ago, johnksss said:

Without the slider, its 1.085V and with slider, it's 1.110V

And my PCIE slot voltage and 16-Pin voltage go no lower than 12.1. While my other card dipped in the 11.8 to 11.9 range.

Please be a good card.😂

 

Banger of a card @johnksss especially if you block it.

 

My Astral is in the neighborhood but your clocks are the chef's kiss! Mine is clocking very nicely, so I know yours is going to go to that next level especially if you block it and shunt/XOC it.

 

 

(FYI, this is my first time punching either boost lock or voltage slider)

 

Stock boost lock:

 

 

voltage slider in "fun times" mode:

 

 

20 minutes ago, tps3443 said:

I really wanted my whole system to be watercooled when I was building this rig and I had kinda prepped with the 4090. We just don’t have any blocks yet for FE models. So, I had already ordered my Fujipoly fullcover rear thermal pad (Originally for the 4090 never used it) but now I’ve ordered some shunts, but I can’t add any of these items until I get a waterblock. FE blocks are taking a little longer than the other models, but they’re coming. I’m really curious what the Bitspower block will even look like. 

 

Do you plan on getting a block though? Are you going to shunt first and run it on the stock HSF? It's gonna kill me seeing you tear it apart especially as I'm sitting here showing my buddy over FT how quiet a FE CAN be with the 5080 FE.

 

 

 

[johnksss]

9 hours ago, tps3443 said:

That’s really good! Which card is this?
 

I would absolutely shunt the card. I went with the 1.5 Ohm 3 watt resistor for my FE from mouser. The stock FE has a single 1 ohm resistor, so stacking 1.5 ohm on top will give me 1,050 watts not counting the PCIe slot power. Bitspower may or may not be sending me a GPU block for testing. Fingers crossed.🤞 I’ve been bugging them to death so we’ll see what happens. 
 

The 5090 even shunted is going to chow down 900+ watts even with lower voltages. These cards are throttling hard though man. The biggest issue is the internal clock that is actually running. The cards are reflecting a higher clock speed externally, but it’s really running much slower than that effectively. It’s like a rubber band effect to prevent crashing due to power limits. So shunting will eliminate this entirely! 
 

 

Can you post some of your benchmarks of before and after showing gains? Shunting has to make sense and actually do something other than show more watts with worse performance. And going by effective clocks. I have yet to be above 3 ghz yet.

 

6 hours ago, jaybee83 said:

 

welp thats the sucky thing, it was can you out of nowhere, even if u had no issues beforehand whatsoever. im happily chugging along on 572.83 while others are crash n burning on it 😄 shouldnt be this way...this is not the way

 

 

really sounds more and more like its driver-related then. was the hard reset overclocking related?

 

 

 

 

 

duuuuude, that card looks crazy! mine only does 1.01V at stock with clocks around 2825 Mhz, with voltage slider maxed it goes up to 1.025V 😄 those voltages are insane! ive only seen maybe 1-2 other cards that can break through 1.1V, u definitely have a special one there. 3000+ Mhz at stock, crazy...

 

and yes ive noticed the same, voltages are all over the place, heavily depending on what ure running and even fluctuate wildly during sessions / bench runs. thats even the case if the power limit hasnt even been reached yet. not sure whats going on, hope this is smth that can be rectified with drivers / vbios updates....

 

overall, its total chaos when it comes to clocks and voltages, every card is so different among 5090s, both in wattage/clocks/voltage, way more variance than with the 4090.

 

those 630W, are those transient spikes or peak load? i might have to give your vbios a go, pls share with the class! 😄 

 

keep us in the loop on your findings, highly interested 🙂 

572.83 had my card not running the same, so on this card i did not use that driver.

5 hours ago, electrosoft said:

 

Banger of a card @johnksss especially if you block it.

 

My Astral is in the neighborhood but your clocks are the chef's kiss! Mine is clocking very nicely, so I know yours is going to go to that next level especially if you block it and shunt/XOC it.

 

 

(FYI, this is my first time punching either boost lock or voltage slider)

 

Stock boost lock:

 

 

voltage slider in "fun times" mode:

 

 

 

Do you plan on getting a block though? Are you going to shunt first and run it on the stock HSF? It's gonna kill me seeing you tear it apart especially as I'm sitting here showing my buddy over FT how quiet a FE CAN be with the 5080 FE.

 

 

 

Not sure if there are any blocks yet. I need to check. Also I remember flashing an Astral vbios and it did not unlock 1.1 volts for me. Now I wonder if that was really an Astral vbios.

[jaybee83]

wondering though if this difference in max reachable voltage actually gives u guys anything extra? especially with that power limit in place... now, where did i put that XOC vbios again....must be somewhere here..... 😛 

[electrosoft]

3 hours ago, johnksss said:

Can you post some of your benchmarks of before and after showing gains? Shunting has to make sense and actually do something other than show more watts with worse performance. And going by effective clocks. I have yet to be above 3 ghz yet.

 

572.83 had my card not running the same, so on this card i did not use that driver.

Not sure if there are any blocks yet. I need to check. Also I remember flashing an Astral vbios and it did not unlock 1.1 volts for me. Now I wonder if that was really an Astral vbios.

 

I know over on the OC forums and 3Dmark, XOC/Shunt modded cards are dominating. For @tps3443, that block is crucial for shunting considering the, ah, tight margins on the FE cooler.

 

Here is my Astral vBIOS. Give it a whirl everybody....hopefully you all get awesome performance out of it.

 

2 hours ago, jaybee83 said:

wondering though if this difference in max reachable voltage actually gives u guys anything extra? especially with that power limit in place... now, where ddid i put that XOC vbios again....must be somewhere here..... 😛 

 

Oh you dropped it. Here you go... 🤣

Astral5090.rom

[tps3443]

2 hours ago, electrosoft said:

 

I know over on the OC forums and 3Dmark, XOC/Shunt modded cards are dominating. For @tps3443, that block is crucial for shunting considering the, ah, tight margins on the FE cooler.

 

Here is my Astral vBIOS. Give it a whirl everybody....hopefully you all get awesome performance out of it.

 

 

Oh you dropped it. Here you go... 🤣

Astral5090.rom 1.87 MB · 0 downloads

 

So I have learned that the GPU voltage is not determined by the bios, so even if we flashed to Astral bios. Some Astrals/Suprim’s cannot hit 1.110v even locked in at 1.100v+ with 100% slider. The voltage is controlled directly by the core. 

13 hours ago, jaybee83 said:

 

welp thats the sucky thing, it can get you out of nowhere, even if u had no issues beforehand whatsoever. im happily chugging along on 572.83 while others are crash n burning on it 😄 shouldnt be this way...this is not the way

 

 

really sounds more and more like its driver-related then. was the hard reset overclocking related?

 

 

 

 

 

duuuuude, that card looks crazy! mine only does 1.01V at stock with clocks around 2825 Mhz, with voltage slider maxed it goes up to 1.025V 😄 those voltages are insane! ive only seen maybe 1-2 other cards that can break through 1.1V, u definitely have a special one there. 3000+ Mhz at stock, crazy...

 

and yes ive noticed the same, voltages are all over the place, heavily depending on what ure running and even fluctuate wildly during sessions / bench runs. thats even the case if the power limit hasnt even been reached yet. not sure whats going on, hope this is smth that can be rectified with drivers / vbios updates....

 

overall, its total chaos when it comes to clocks and voltages, every card is so different among 5090s, both in wattage/clocks/voltage, way more variance than with the 4090.

 

those 630W, are those transient spikes or peak load? i might have to give your vbios a go, pls share with the class! 😄 

 

keep us in the loop on your findings, highly interested 🙂 

I will never see a 1075 in real world either mine is typically 0.990-1.010 and with 100% slider it’ll show a 1.020 So, in order to see the 1075v “At Idle” you need to set 100% voltage slider, and set the 1.110v point on the graph and hit Shift+L to lock it in. This will lock in your highest GPU’s voltage “That it needs” for that set frequency at the 1.1100 voltage graph point. This voltage isn’t controlled by the bios. It’s completely based on the cores needs. 

 

12 hours ago, electrosoft said:

 

Banger of a card @johnksss especially if you block it.

 

My Astral is in the neighborhood but your clocks are the chef's kiss! Mine is clocking very nicely, so I know yours is going to go to that next level especially if you block it and shunt/XOC it.

 

 

(FYI, this is my first time punching either boost lock or voltage slider)

 

Stock boost lock:

 

 

voltage slider in "fun times" mode:

 

 

 

Do you plan on getting a block though? Are you going to shunt first and run it on the stock HSF? It's gonna kill me seeing you tear it apart especially as I'm sitting here showing my buddy over FT how quiet a FE CAN be with the 5080 FE.

 

 

 

My 5090FE is very quiet. 575w in a closed case it stays below 50% Auto fan speed. Undervolting it drops that to sub 40% fan speeds. The main thing with my card is I never see power usage this high unless I’m running Furmark or 4K scaled up 200% to 8K. Since I’m shunt modding, I have to remove the cooler which means I won’t be reinstalling the original air cooler. I’m not taking the cooler off until I have a waterblock in hand though. 

6 hours ago, johnksss said:

Can you post some of your benchmarks of before and after showing gains? Shunting has to make sense and actually do something other than show more watts with worse performance. And going by effective clocks. I have yet to be above 3 ghz yet.

 

572.83 had my card not running the same, so on this card i did not use that driver.

Not sure if there are any blocks yet. I need to check. Also I remember flashing an Astral vbios and it did not unlock 1.1 volts for me. Now I wonder if that was really an Astral vbios.

Once I get a block so I can tear my card apart, I’m going to solder on the shunt. This will make a massive difference for all 5090’s in their internal effective clock speed.

[electrosoft]

1 hour ago, tps3443 said:

 

So I have learned that the GPU voltage is not determined by the bios, so even if we flashed to Astral bios. Some Astrals/Suprim’s cannot hit 1.110v even locked in at 1.100v+ with 100% slider. The voltage is controlled directly by the core.

 

 

Good find!

 

So silicon lottery is in play again as always.... 😞

If someone is dedicated enough I can see binning cores and memory (if we're ever allowed to exceed +3000) to compete. I mean, I'm clearly not sad for my Astral but still.

 

I think we always have to remember all that is promised out of the box is the advertised performance...

 

....we also have to remember most products also come with a 15-30 day return policy or you can sell them off (like CPUs) to friends/family/clients who will never, ever overclock and a 5090 is a 5090 is a 5090 to them.

 

1 hour ago, tps3443 said:

 

I will never see a 1075 in real world either mine is typically 0.990-1.010 and with 100% slider it’ll show a 1.020 So, in order to see the 1075v “At Idle” you need to set 100% voltage slider, and set the 1.110v point on the graph and hit Shift+L to lock it in. This will lock in your highest GPU’s voltage “That it needs” for that set frequency at the 1.1100 voltage graph point. This voltage isn’t controlled by the bios. It’s completely based on the cores needs. 

 

 

What does your FE say boost locked with PX1 w/ and w/o 100% slider?

 

1 hour ago, tps3443 said:

Once I get a block so I can tear my card apart, I’m going to solder on the shunt. This will make a massive difference for all 5090’s in their internal effective clock speed.

 

For me, it comes down to coil whine. I'd have to hear it at >200fps+ and TS GT2 run to see what it's working with coil whine wise. 50% fans on these FEs is definitely audible but then again I'm running an open case.

 

1 hour ago, tps3443 said:

Once I get a block so I can tear my card apart, I’m going to solder on the shunt. This will make a massive difference for all 5090’s in their internal effective clock speed.

 

True, even the most average FE should beat a good sample 5090 if it is blocked and shunted. 🙂

 

Looking forward to before and after results.

[jaybee83]

Just now, electrosoft said:

Good find!

 

So silicon lottery is in play again as always.... 😞

If someone is dedicated enough I can see binning cores and memory (if we're ever allowed to exceed +3000) to compete. I mean, I'm clearly not sad for my Astral but still.

 

I think we always have to remember all that is promised out of the box is the advertised performance...

 

....we also have to remember most products also come with a 15-30 day return policy or you can sell them off (like CPUs) to friends/family/clients who will never, ever overclock and a 5090 is a 5090 is a 5090 to them.

 

What does your FE say boost locked with PX1 w/ and w/o 100% slider?

 

For me, it comes down to coil whine. I'd have to hear it at >200fps+ and TS GT2 run to see what it's working with coil whine wise. 50% fans on these FEs is definitely audible but then again I'm running an open case.

 

True, even the most average FE should beat a good sample 5090 if it is blocked and shunted. 🙂

 

Looking forward to before and after results.

 

no dice on unlocked vRAM OC, that is hardlocked in the vBios. so well need the XOC vBios including that special sauce software to drive it...

and yeah, can confirm from what ive seen, voltages of a card dont change when flashing different vbios versions. however, there might be different boosting behaviour in terms of power spikes that the card is allowed to do.

[johnksss]

@tps3443

I was thinking it was already done. My bad.

 

As to voltage, you can try this and see if it will lock the higher voltage. You are not trying to break any records or anything, only trying to lock the voltage and move the core. 

In the picture it's only 1105 because I'm not using a full screen bench test, so it slightly lowers itself. That and temps. Also that yellow line is the locked voltage point.

[tps3443]

19 minutes ago, johnksss said:

@tps3443

I was thinking it was already done. My bad.

 

As to voltage, you can try this and see if it will lock the higher voltage. You are not trying to break any records or anything, only trying to lock the voltage and move the core. 

In the picture it's only 1105 because I'm not using a full screen bench test, so it slightly lowers itself. That and temps. Also that yellow line is the locked voltage point.

Also too, when checking voltage. Make sure your temp is below 56c. That is some vital point with 5090’s. So I’d test yours again, just like you’re doing now only with sub 56c temps. This will reflect its true max voltage. Or at least that’s how it was explained to me. 

[johnksss]

2 minutes ago, tps3443 said:

Also too, when checking voltage. Make sure your temp is below 56c. That is some vital point with 5090’s. So I’d test yours again, just like you’re doing now only with sub 56c temps. This will reflect its true max voltage. Or at least that’s how it was explained to me. 

Oh, I was only doing that to show you how the locked voltage worked. To see if you could lock at 1.V. I wasn't trying to do anything. LOL

[tps3443]

32 minutes ago, johnksss said:

Oh, I was only doing that to show you how the locked voltage worked. To see if you could lock at 1.V. I wasn't trying to do anything. LOL

I cannot lock that voltage. My card maxes out even with lower temps.

 

[tps3443]

45 minutes ago, johnksss said:

Oh, I was only doing that to show you how the locked voltage worked. To see if you could lock at 1.V. I wasn't trying to do anything. LOL

Yzonker has a 5090 Astral shunt modded and on water and he cannot lock more than 1075mv. So it really seems to be all over the place card to card.