i7 840qm vs 920xm

considering overall parameters (clock speed, price, TDP, idle and load temps, etc.), would the 840qm be better than the 920xm? is it wise to downgrade from a 920xm to an 840qm?

 

Looks like two of the options on the X8100 I ordered earlier this month, only answer I can give is that you lose the ability to overclock on the 920xm and some speed without turboboost enabled.

 

Well, I really don't need overclocking since I'm in a tropical region and have very little headroom for increased temps. With regards to the speed without turboboost enabled, is it that significant?

 

See this thread, last post on the 1st page for the exact difference.

For practical purposes, unless you are a photoshop or video encoding fiend, there's almost no application other than artificial benchmarks that will demonstrate a clear performance difference between the two processors.

 

The big thing that seems to go unmentioned, is the TDP difference. With the 840's 45TDP, you're not going to get any turbo boost on 8 threads, and very little with the other loads.

But like Harleyquin said, if you don't do a lot of media encoding, emulating, compression or heavy multitasking, you won't notice a huge difference.

 

Ah. So the 840's turboboost capability is limited by its TDP. I thought both can turboost 8 threads equally.

 

All mobile i7/i5's are limited by TDP. That's the purpose of turbo boost, you get as much speed as it can give you without going over the power limit. Running 8 threads uses a lot of power. Quads are 45w, duals are 35w, and xm's are 55w(adjustable).

 

Ok. How much do you think is a reasonable price (high side) to sell off my 920XM? Sorry if this is the wrong section to ask, but I hope I could ask for your opinion anyway.

 

Couple of things, keep in mind even though the 840QM is rated 45W and the 920XM at 55W and since they have very similar clock speeds the 840QM really is going to have very close to the same real world TDP as the 920XM

So you want to sell a 920XM to buy an 840QM? Is your 920 OEM or ES? If it’s OEM you are looking in the $400-600 at the most (just looking at what is available on ebay) ES are fetching high $300’s.

To me your better off keeping the 920XM

 

It's OEM. Ok I got that. If I could make money in selling the 920XM and buying the 840QM, I would go for that. Else, I won't go that way.

 

i suggest u keep ur 920xm.. it way better and ocing still helps... also massive turbo is possible on 8 threads if u have an M17x... ur choice.. personally , iw would keep 920xm.

 

I don't think it quite works that way. Turbo boost will only be bound by temperature and the available turbo boost multipliers. TDP means nothing to it, not that TDP means much about anything. They would not have turbo boost multipliers if the processor wouldn't do them.

While Intel hints at the details of turbo boost, they are all hidden. They used to say that the processor measures current, power consumption, temperature, and the number of active cores to determine if it can turbo boost up to the next level, remain where it is, or turn down the turbo boost. Currently, they say that they estimate if the current is not too high or power consumption is not too high. Good thing they changed it since the processor can't measure it. But these are things that they can estimate before they even make the chip, building-in the constraints already by determining operating voltage and frequencies. I'm almost certain this part is just mumbo-jumbo made to look like turbo-boost is cooler than it really is and that the only factors during use are an upper temperature bound, number of active cores, and available multipliers.

As far as turbo-boost being cooler than it really is, the technology is absolutely nothing great nor nothing new.

Also if you have 4 threads it will turbo boost 4 cores, not 2 cores. 2 cores with hyperthreading are way less powerful than 4 cores without.

Don't worry about your climate. Unless the difference is living in an igloo or living in a tent in the sahara, there shouldn't be big difference in overclocking.

 

Keep the 920XM. I have OCed it and ran all four cores and 8 threads at 3.8GHz. As long as you can keep temps down, the 920XM can run almost 100% faster than the 840M. All you need is to use the "throttle stop" program to increase your TDP/TDC and you can raise the power of which the CPU has turbo boost

 

Well, what if as of the moment I'm already reaching around 90C when I stress test my 920XM? Considering that I already use Shin Estus X23 thermal paste (one of the best pastes in the market)? Do I still have a leeway to overclock?

 

What is your laptop? 90C is getting up there, You really have no leeway with temps that high at stock settings.

 

What are ambient temps? And like othonda asked, what is your laptop?

 

I'm positive it works this way. Using throttlestop, you can test all of this out. For example: My old 820QM would peak around 81*C running 8 threads in prime, but would receive zero turbo boost because it was over the TDP limit. Running 4 threads, I got some turbo, but not the max and temps were about the same. I tested and my 940XM does the same thing if you lower the TDP to 45.

If you raise the TDP, it won't throttle, even at higher temps until it reaches the power output max. Revelator(sp?) did a chart showing the effect of TDP/ratio changes on wprime times. The higher TDP, the better the score, even at the same ratios.

Heat is a product of clock rate, so the two are connected, but I think turbo is dependent on power and not heat. Of course, if the notebook is made to throttle at higher temps, that's a different story.

 

Soviet did a study a while back that he posted in the Sager 8760 that showed the turbo clock was throttling based on certain temp threasholds, so I think this is a bit more complicated than that. So long as your temps are below this threshold the throttlestop turbo boast and TDP boast will function, But my guess is that once you exceed the temp threasholds your probably forced to throttle down.

 

Ok, thanks for the info. I have a desktop i7 860 and it would turbo boost no matter what, even if already overclocked to ridiculous levels. I can't believe it's not exceeding the 95w TDP at full load when at over 4Ghz to begin with. Of course I have that shut off, but I was just playing with it a bit.

 

If TDP is still the controlling factor, then whether it's power or heat that triggers it isn't really important to the main issue. The 45TDP of the QM limits their turbo under most situations. At 80TDP I can maintain 95*C and have full consistent turbo. The reason I believe it's more about power and not heat, is that on cold days. (An ambient difference of 10-15*C with the window open) It still throttles in the same places even though the core temps are much lower.

In the end, higher TDP means better performance. An 840QM with 55TDP is going to perform better than an 840QM with a 45TDP.

 

Well, you could always sell your 920XM and get a nicce 920XM ES and make like 230 bucks...

personally, I would keep the 920XM.

 

Correct me if I am wrong but I think they have the W860CU with the GTX 285M GPU. If that is the case, perhaps the CPU vent sticker is still intact. My 820QM never goes beyond 71C under load and I play a lot of GTA IV. However, when I still had the CPU vent sticker on, my CPU temps under load were around 80C.

 

I was just saying that I could hit 95 with a really high TDP setting. I don't run that high 24/7, just when testing. I have the alarm function in Throttlestop set to 92, so it never goes over that. Usually I get 45-50 idle and 80-85 load at my current settings.

 

your getting this on stock?
did you try reapplying your paste, just to make sure the application is done well?

 

clevo w860cu
920xm oem
gtx 285m
cpu sticker removed

i already had two w860cu's. i've tried shin etsu x23 and mx2. same results. i know how to apply thermal paste.

4 cores of cpu idles at around 53c.. under load, maxes at 90c.

ambient temp is 35c.

 

This is incorrect thinking.

Current is measured via the VRM current sense output IMON. AFAIK the CPU doesn't measure VCore but since it does request the voltage by VID then if say it requested 1.000V and current IMON indicates 50A then power could be calculated by the CPU as 50W. Exceeding the rated TDP will turn off turbo.

Turbo ratios will depend on active cores if the ratio limits are different across the number of cores. For instance with your 860 all 4 cores can use the 22x turbo bin. With only 2 cores active that goes up to 25x.

Probably the reason you don't see anything on your desktop board might be due to some desktop boards using a fixed current for IMON instead of the true current. For example while your 860 might be drawing 60A the CPU is being told it is only drawing 20A. Since the current effects both TDC and TDP then by using this workaround you will never see TDP/TDC throttling. I suspect your desktop board works using the fixed current reporting since it would explain your earlier thinking on TDP. AFAIK the same hardware technique should be able to be applied to mobile chips if TDP/TDC limits are worrisome.

 

gonna have to say...the 840 is going to take a loss at pretty much everything...if you tested them both out...you would already know....
and since you can get an oem off of ebay for about 600 (920xm)....it doesn't have price in it's corner either.

 

so if you guys have a choice between the 840qm and 920xm for the same price, most probably the 920xm would be likely the wise choice?

 

yes.
a 920 can run almost exactly like a 840, but an 840 can never ever run like a 920.