"Freak of Video Cards"

[f1sherman]

how is it 13% less?

q ~ T^4

()

 

[pantsaregood]

q ~ T^4

()

Thank you.

Describing 10 C as twice as hot as 5 C is fundamentally wrong. 10 K is twice as hot as 5 K. 5 C is not 50% as hot as 10 C.

 

[tynopik]

If today we were in the Sahara desert and the daily peak temperature reached 80*C, but then subsided to 68*C towards the evening, the drop in the temperature would amount to a 15% reduction in temperature.

80C = 176F
68C = 154.4F

Therefore using the true imperial scale we can see that there was in fact only a 12.3% reduction in temperature.

 

[lehtv]

They should be comparing the observed temperatures to their ambient temperature, that way their percentages would actually make sense from an air cooling point of view.

Assuming their ambient temperature is 25C, for instance, their observed load temperatures are 65ΔT for the reference card and 53ΔT for the Gigabyte SOC. That's an 18% reduction in ΔT, or an 18% improvement in air cooling efficiency.

 

[f1sherman]

Describing 10 C as twice as hot as 5 C is fundamentally wrong.

Good point
Now stop right there. Don't blow it now!

I kid I kid...

Anyway you might have been wrong before, but you got us thinking, so ty for that.

 

[KingFatty]

If we ignore the units, and just talk about numbers, can you even use percentage to relate to numbers?

Like can you say the number changed by x%? Maybe that's what the reviewers do, they just talk about how a number can change by a percentage, forgetting about the units of temperature.

 

[RussianSensation]

80C = 176F
68C = 154.4F

Therefore using the true imperial scale we can see that there was in fact only a 12.3% reduction in temperature.

I don't live in the US. The Metric System is the "International System of Units" - the official system of measurement in almost every country in the world. The United States is the only or one of the few industrialized country that does not use the metric system as its official system of measurement. Sorry, but I am going to go with 200+ other countries in the world as being correct than 1-2 countries that continue to use pounds, feet and Fahrenheit.

I took physics and I realize to calculate "proper" % change in temperature you have to use Absolute/Kelvin temperatures. If I go outside and the temperature drops from 32*C to 23*C, 99% of the world would call that a 9*C drop in temperatures and not convert it to any F or Absolute Kelvin scale. Let's not make this more complicated than it is. The cooler is more effective than the stock cooler and runs quieter. Using proper physics, you and I are both incorrect since we didn't use the Absolute scale anyway hehe.

If we ignore the units, and just talk about numbers, can you even use percentage to relate to numbers?

Like can you say the number changed by x%? Maybe that's what the reviewers do, they just talk about how a number can change by a percentage, forgetting about the units of temperature.

That's because they use temperature as the unit of measure, not using the proper physics definition. They are just focusing on the #s which is why their % change is purely mathematics based, not physics based. For example, if you weigh 90 KGs and you lose 20 KGs of fat through exercise, your new weight is 70 KGs. You would say:

1) You lost 20 KGs of weight;
2) Your new weight is 77.78% of your old weight (70 / 90 * 100%);
3) You have experienced a -22.22% change in weight.

When ignoring the units of measure as you specified, one can calculate the % change using the simple math formula (New - Old) / Old * 100%. However, for temperatures, this isn't entirely accurate based on the physics definition.

 

[lehtv]

If we ignore the units, and just talk about numbers, can you even use percentage to relate to numbers?

Sure, why not...? Two is 100% more than one. Nine is 50% more than six. Six is 33% less than nine.

Like can you say the number changed by x%?

Well that you can't do. Percentual change is directional so you need to indicate whether it's more or less. Saying that two numbers have a difference of x % is meaningless, although often it will be obvious whether the author is talking about a decrease or an increase.

Maybe that's what the reviewers do, they just talk about how a number can change by a percentage, forgetting about the units of temperature.

Indeed, that's the problem.

 

[Smoblikat]

Interesting take on mathematics lol! By my reckoning 10% of 85 is 8.5.
20% of 85 =17 (2x8.5). 85 minus 17 =............68!
68 is 80% of 85. So a 20% reduction seems correct although I wasn't terribly good at maths at school. The idea that 68 is 97% of 85 is so glaringly obviously wrong I'm surprised you could even write it. My maths is based on the assumption that a 20% reduction in the number of 100 idiots is 80 idiots.

Lol, I agree with you but I think I know what the other guy was saying. I think he was converting C to F and getting numbers based off of that. So your units of measurment were wrong.

 

[chimaxi83]

I laugh at the Kelvin nazis everytime they try and correct people here. Almost as bad as the grammar nazis. He's not the only one here, either :sneaky: As if computer hardware reviewers are supposed to start using a scale reserved almost elusively to those working in a scientific field.

 

[RussianSensation]

Lol, I agree with you but I think I know what the other guy was saying. I think he was converting C to F and getting numbers based off of that. So your units of measurment were wrong.

Celsius > Fahrenheit. The world =! USA. Trying to tell people they are wrong because they are using the Metric system instead of Imperial is strange since it's the International system for units of measure. The proper temperature change uses Kelvin anyway, so both C/F are both equally wrong without an adjustment for that physics definition. :biggrin:

 

[Shmee]

haha, highly technical is ===>

but srsly, I think K and C make the most sense to me, but they just have different starting points.

 

[tynopik]

Sorry, but I am going to go with 200+ other countries in the world as being correct than 1-2 countries that continue to use pounds, feet and Fahrenheit.

It was to point out the absurdity of your methodology. Why should the %temp drop depend on which scale you're using?

If a long jumper goes 10% farther, that's true whether you're using inches or cm.

If a car goes 10% faster, that's true whether you're using kph or mph.

If a parcel of land is 10% larger, that's true whether you're using acres or hectares.

If a pool is 10% bigger, that's true whether you're using gallons or liters.

The fact that your method results in different answers depending on which scale you're using should clue you in that you're doing it wrong.

If I go outside and the temperature drops from 32*C to 23*C, 99% of the world would call that a 9*C drop in temperatures and not convert it to any F or Absolute Kelvin scale.

Saying it was a 9 DEGREE drop is perfectly legitimate

Saying it was a 28 PERCENT drop is a bunch of hooey

If you really want to use percentages, lehtv showed the way:

They should be comparing the observed temperatures to their ambient temperature, that way their percentages would actually make sense from an air cooling point of view.

Assuming their ambient temperature is 25C, for instance, their observed load temperatures are 65ΔT for the reference card and 53ΔT for the Gigabyte SOC. That's an 18% reduction in ΔT, or an 18% improvement in air cooling efficiency.

25C = 77F
78C = 172.4F
90C = 117F

95.4ΔT for the Gigabyte SOC
117ΔT for the reference card

(117-95.4)/117 = 18% reduction in ΔT

see? same answer no matter which scale you use

lehtv wins one internet

However, for temperatures, this is COMPLETELY NONSENSICAL based on the physics definition.

FTFY

 

[AtenRa]

It was to point out the absurdity of your methodology. Why should the %temp drop depend on which scale you're using?

If a long jumper goes 10% farther, that's true whether you're using inches or cm.

If a car goes 10% faster, that's true whether you're using kph or mph.

If a parcel of land is 10% larger, that's true whether you're using acres or hectares.

If a pool is 10% bigger, that's true whether you're using gallons or liters.

The fact that your method results in different answers depending on which scale you're using should clue you in that you're doing it wrong.

All of the above never have a negative equivalent relatively to each other.

1 cm = 0.4 in
0.5 cm = 0.2 in
0.005 cm = 0.002 in

and so on

1 kph = 0.62 mph
0.5 kph = 0.31 mph
0.005 kph = 0.0031 mph

And so on

The problem with Fahrenheit is that its equivalent in Celsius is NEGATIVE, making it impossible to have a direct 100% percentage calculation.

1F = -17,2222C
0.5F = -17,5C

and so on

Saying it was a 9 DEGREE drop is perfectly legitimate

Saying it was a 28 PERCENT drop is a bunch of hooey

If you really want to use percentages, lehtv showed the way:

32c to 23c is a 9 degrees CELSIUS drop and YES it is a 28% PERCENTAGE drop in the Celsius scale.

If we use Fahrenheit units then it is not 28% percentage drop in Fahrenheit units.

Hope i didnt write anything stupid at 2:00 am

 

[pantsaregood]

To the guy who said I converted C to F: No, I converted C to K. I never used Fahrenheit. I never explicitly stated I was using Kelvin, but my offsetting the Celsius scale by 273 degrees would clue most people in on that.

You could technically use Rankine as well. But no one does that.

It isn't right to claim Fahrenheit has a problem because of negative temperatures. Both Fahrenheit and Celsius use arbitrary zeroes, so why is one more right than the other? That's the issue with relative scales. They're arbitrary.

Also, when a reviewer is making statistical claims, they need to be valid. These simply aren't.

 

[tynopik]

YES it is a 28% PERCENTAGE drop in the Celsius scale

It is a 28% drop in temperature RELATIVE TO THE FREEZING POINT OF WATER. That is true regardless of whether you're using Celsius or Fahrenheit.

Just as lehtv showed how to normalize to room temperature, you could normalize both to the freezing point of water instead and then they would both say 28%

It has NOTHING do with Celsius or Kelvin or Fahrenheit or Rankine and EVERYTHING to do with what you choose as your baseline

 

[blackened23]

You guys sure are slapping each other a lot over this temp thing....only powermad video cards with torture chamber cooling are slap worthy.

 

[SirPauly]

Hehe, debating for the sake of debating!

Interesting sku!

 

[reallyscrued]

This thread taught me more math than my public education.

 

[Siberian]

wouldn't this thing still get smoked by a 690? I'm not sure of the point ofc it.

 

[pantsaregood]

wouldn't this thing still get smoked by a 690? I'm not sure of the point ofc it.

yes
a high-end single-gpu card would absolutely get smoked by a dual-gpu card

gigabyte will then produce the hd 7990 soc
it will be 18 inches long, take four expansion slots, and require a constant supply of liquid helium

 

[3DVagabond]

Just thought I'd add this little piece of kit.

It's from a review at Tom's. A five slot cooler! Why would anyone buy this?

 

[lehtv]

Imagine a Gigabyte 7970 GHz Edition X2 SOC. 600 Watts power consumption.

 

[moonbogg]

Kinda ridiculous.

 

[Elfear]

wouldn't this thing still get smoked by a 690? I'm not sure of the point ofc it.

$530 vs $1000.

Can't really compare the two with the huge price discrepancy. You could almost buy two SOCs for the price of a 690.