NVIDIA GTX 1650 GDDR6 vs. GDDR5 Benchmark: Big Uplift in Performance

by birtanpublished on September 15, 2020

Nvidia says that the gddr5 supply is running out and in accordance with that the company is refreshing some of its active GG r5 products with GG dr six to keep them going this isn't a common occurrence but it does happen when there's a brewing switchover between two

Memory generations we also see this happen for other SM DS or even thermal pads when supply of one dries up and a new supply has to be sourced for an active SKU that's why most manufacturers don't define the individual components

That are on there PCBs in this instance GD r 6 has potential to influence performance in a major potentially positive direction but Nvidia's reference GPU clocks for the gtx 1650 g6 have also been adjusted downward this

Might negate some of the impacts and so for both power consumption and for game performance reasons we're going to be looking at the gddr5 and g2 dr6 versions of the 1650 today including a clock for clock comparison on the GPU core before

That this video is brought to you by gigabyte – or srt x20 atti extreme the r-tx 2080 ti extreme is built with a triple axial cooling solution and ready for anyone interested in intermediate GPU overclocking although it's also up

For gaming out of the box the gigabyte 2080 TI can reach the higher performance range required to play games at frame rates at and beyond 144 FPS couples particularly well with games like Call of Duty warzone Rainbow six siege and

Other competitive FP assets gigabytes extreme is built to be a looker for system builders going for extra visual flair learn more at the link in the description below this content is going

To be really simple we're not going to include all the other cards in our benchmarks for this one we obviously have other data but you can look at previous benchmarks for that this is so focused solely on these 1650 cards and

Just the non super SKUs so one of these is g6 one is g5 there otherwise the same we bought an EVGA SKU from new egg and our previous SKU is a GD r5 EVGA SC ultra from the original review cycle the biggest question here again is to set

The clocks equal to each other that's pretty easy and we'll show you a frequency plot in the first set of charts to illustrate what we did to do that and separately the next thing I look at other than just g5 versus g6

Performing on the same GPU obviously stock the stock matters too because most people aren't going to normalize the clocks they're going to plug it in and forget about it

Pricing is about the same the original card the gddr5 1650 which we said was dead on arrival was priced at $170 at the time we reviewed it and it's still about that price it's 150 to 170 on average for a g5 version the GDD r6

Version is about 160 to 170 dollars it has improved in performance the 1650 super was a pretty good card for the price the 1650 was not well positioned at launch like we said it was dead on arrival and it remains poorly positioned

Once the 1650 super came out and further buried it so it's still kind of where it was before that GD r6 version it's just that it's a bit better now but we don't necessarily recommend buying it actually we don't really recommend buying it if

You can spend a little bit more for something else anyway that's a discussion you can watch in our GTX 1650 super review for this one we'll talk about these two cards specifically there's a couple of other things to

Cover here but we'll throw that in the conclusion so let's just get straight into the benchmark since this one's so simple we need to first establish the frequency at which the cards run because any like-for-like comparison will be

Made more difficult by different core clocks we're not intending to compare the core clock change here we want to compare the memory change so equalizing for frequency is a must we'll also have benchmarks with the

Stock to stock card performance of course just because that's probably how most people will be using it we're using times by gt1 and gt2 for this frequency charge with a gap in between them the gddr5 original model ran at about 1950

Mega Hertz flat for the GPU core frequency it started out 1980 and then dropped to 1915 for the rest of GT want for GT to the gddr5 model ran at about 19 50 megahertz when stock with a dip to 1930 towards the end running stock the

GDD r6 model has more trouble holding a hard-line frequency but tends to hold that about 1905 megahertz that 45 mega Hertz deficit is important and needs to be corrected for in any like-for-like comparison with just the memory change

Being tested with a 40 mega Hertz offset and a +3 percent increase in power budget we were able to hit the frequencies and manually align them the GDD are six cards extremely slight overclock if you can even call it

That at these numbers held about 1950 megahertz it's still plus or minus ten megahertz peak to peak but overall it's much closer and close enough that we can call them about the same for this one-off

Test 3dmark times pi gives us some useful scoring that can be more reliable in establishing deltas in the memory performance gt1 and gt2 are split out in a way that we can better understand what each test is doing they isolate for

Memory or for course so that's useful in a synthetic fashion for this one the gtx 1650 with gddr5 scored 19.6 FPS GG to average one stock Wiley stock g6 card with its core frequency deficit included scored 22 point 2 FPS and that's an

Increase of 13.3% from the g5 to the g6 card that's without even equalizing the frequency gt1 posted a difference of 21.5 to twenty four point nine or a fifteen point six percent difference although not shown here because it blow

Out the scale the scoring was 33 61 points versus 38 51 points on the g6 card if you care or 3900 points with the g6 card with the core clocks equalized the gt-r six cards GPU frequency set approximately the same about nineteen

Fifty megahertz resulted in twenty two point five one FPS for gt2 and twenty five point two for FPS for GT want as a note every fraction of a frame in time spy is important to scoring it's not like a game where it'd be totally

Irrelevant it's harder to gain those fractions of a frame in this benchmark than in game benchmarks the GDD are six o´clock controlled test puts it at a massive 17.2% ahead of the g5 card at the same clocks but we should test some

Other games to get a more real-world representation for f1 at 1080p the stock gddr5 test ran 62 FPS average 35 FPS 1% low and 18.4 FPS 0.1% lows representing the frame times the stock GTR six-card improved this result by about 5% while

Running at lower core clocks demonstrating the benefit of the GDD r6 memory upgrade matching the core clocks together pushes that by an additional 3.5% it's a 67.5 FPS average with the low results

Within test error the improvement at roughly the same core clock but with GG dr six is about 9% over the original gddr5 car at 1440p that difference is about eight point eight percent between the matched core results and the stock

Gddr5 result at fifty point six FPS average to forty six point five FPS average the GDD are six stock card ran at forty eight point eight FPS average maintaining a rough five percent advantage as seen at 1080p

We run multiple resolutions to see if there's any scaling across them as the workload shifts so next we'll look at 4k despite it being completely unreasonable for these cards at 4k the improvement continues to shrink

We're down from nine percent at 1080p it's a 8.2 percent at 4k measured between the gddr5 stock card and the GD r 6 card with the match score frequency the GD r 6 card is about four point nine percent ahead so not much has really

Changed here out over the tomb raider is next to starting again at 1080p for this test the gtx 1650 with gddr5 ran at 52 FPS average versus 56 FPS on the g 6 stock model that's an increase of 8.7 percent and that's in spite of the

Decrease in core clocks on the g6 stock model matching the frequency got it 257 FPS average so not much change and within run to run variation of the stock g6 test the matched core difference ends up being about nine point eight percent

From g5 g six which is similar to the nine percent improvement we saw in the previous game benchmark moving on to 1440p to check for scaling the match score difference between g5 and g6 ends up at about eleven percent so it has

Increased with the higher resolution that makes sense and memory intensive scenarios like this one if you were to just buy the card and use it out of box like most people would do the difference is still 9.3 percent between the thirty

Seven point seven FPS and thirty four point five FPS average results and we didn't run this one in 4k because it became too stuttery even for a synthetic analysis GTA 5 showed a bigger change we retested this one to validate and count

With the same results at 1080p the gtx 1650 g5 card scored almost exactly the same as it did our initial review illustrating why we like keeping GTA around on the bench it's consistent and it's reliable even

Across five years of testing that makes it easier to fact-check against older data for scaling the clock matched 1650 g6 held 71 FPS average a markedly dat 12.5% that's the biggest in-game increase we've yet monitored the stock

To stock test also showed an improvement at 70 point 6 FPS average versus the original 63 although not shown on this charge just for reference the 1650 supercard runs this game at about 83 FPS average so the gap has narrowed between

The 1650s Q's at 1440p the 1650 g5 the g6 gap increases to 14 point 4 percent on the matched core result or about eleven point four percent from stock to stock or out of box performance the increase in resolution has increased the

Bandwidth requirement and that's reflected positively with the jump to GD dr6 we'd expect that continued with 4k as the bandwidth requirement increases even further in this instance it's at 15.7% improved from gddr5 to G GD r6

With matched core clocks or 12.9% if you look stock to stock where the g6 card is running at a slower core frequency we know no one is going to play with these settings or at these frame rates but the point is to illustrate the scaling under

Heavier memory load conditions and under higher strain on the bus the next title is hitman two with dx12 enabled for this one we get the opposite not every game cares as much about memory bandwidth and hitman two is one of them in this title

We're more limited by the reduced SM account than we are by the memory at least when benchmarks with these settings the GTX 1650s are maximally 5.4% apart back to the original numbers we saw in f1 1440p shows the same thing

The maximum top-to-bottom difference is 5.5 percent functionally unchanging from the 1080p deltas this is a game that doesn't care much about the memory bandwidth between these specific cars doesn't mean it doesn't care in general

About memory bandwidth just between these 1650 models where they're going to be limited by something else in the core first just to really illustrate that point even 4k proves this we measured a 5.4 percent Delta top-to-bottom at 4k

Resolution and stock the stock results aren't that different either this is a big difference from what we saw in GTA 5 just for sanity we also ran hitman 2 with DirectX 11 in this version of the

Benchmark the results posted an 8.5% increase in performance from the gddr5 1650 to the g6 1650 the two GTR 6 cards were about the same given the small change in clocks but the memory difference still benefited the new card

In a measurable way although maybe not a meaningful one for purely synthetic reasons we also included in ashes of the singularity escalation run at 4k this is a DirectX 12 benchmark as are some of the previous tests and it's one that is

Useful for restraining cards kind of like thymus bias in this one our maximum difference was at nine point one percent from top to bottom at twenty eight point two FPS average versus twenty five point nine FPS average for the g5 card both at

The same clocks the stock to stock comparison still plots a seven percent performance increase from the memory even with the clock deficit on the core the last chart is a quick power test we took power consumption at the PCIe slot

And at the PCIe 12 volt cables with a custom interposer and with Hardware monitoring the stock gtx 1650 with gddr5 to about 84 watts and fir mark compared to about eighty six point four watts with the g6 card this is within

Tolerance of potential error or test variants there's about 91 watts when they were a clock matched so there is an increase the time spy numbers were about 10 watts different between the two cards more interestingly we noticed that the

PCIe slot only ran about 12 watts for the 12 volt lines when tested for the gddr5 card while the power was balanced evenly with the PCIe slot and the cable for the gt-r six card it was a difference of one amp on the original to

A split of three point four to three point eight amps on each of the PCIe slot and cable for the newer card Nvidia as well within the slot spec but it still saw a power consumption increase on the PCIe slot and they decrease on

The cables to accommodate the newer memory it seems okay so that'll recover for those the most interesting finding here although not particularly useful in any way is just that the power consumption seems to be more evenly

Split between PCIe on the motherboard and PCIe in the cables and we 12 volt lines on the PCIe slot and then obviously the the power connector so that doesn't have any implications of

Anything it just means that it's more evenly distributed why we're not really sure it's something with how the power delivery set up but it's not exceeding the power consumption limitations and either connector so and it's not really

That different in power consumption than originally at most preceding 10 watts difference in some workloads so it's all that different in in any power consumption metric it's better in performance it's in some cases a

Significant uplift we had a couple of other games to that we tested where we're seeing some of the higher numbers like we saw towards the 14 percent marker but that's about the most you can expect out of it so a big jump yes and

Definitely worth the upgrade if you had to buy it new today obviously shouldn't be upgrading near 1650 tgr5 model to a g6 model that's why it's the money also in general we don't think you should really be buying a gtx 1650 it's not

It's not an objectively bad card in that it does not have any flaws in basic operation but it is a comparatively bad card in that it's a bad deal when you look at even on videos on offering 1650 Super's are much more worth it even with

The performance increase that these got and the reason that video brought the clocks down a bit was to try and mitigate that performance increase so that they didn't cannibalize their own skew the closest AMD competitor we the

5500 XT or something like that but that one had a bit of a troubled launch so yeah it is improved but otherwise remains about where we thought if you're gonna buy one then definitely get the gt-r sixth version don't buy the GG r5

Version it is a meaningful improvement you shouldn't just kind of brush it off as a small change it's a big change here and in some games it's only 5% but and get them out of them we've seen 10% or so or more there is one complaint I want

To point out because we're not going to do a separate teardown on these but the EVGA card at least the one that we picked up it does not have any memory cooling whatsoever so evj says it technically operates in spec

We believe them it probably does technically operate within spec but GTR 6 is pretty hot and even just it's just like they have other coolers it's a down costing thing but this is one of those things where as reviewers who work with

A lot of cards we don't really feel comfortable down costing in some areas and one of those would be memory cooling because long-term use you don't know what kind of system the user is putting it in a lot of people have hot cases

They run hot configurations because they're not really built in a way that's mind full of thermals and now you have absolutely nothing to fall back on for the memory except for the fan and here's the problem with a GPU that's well

Cooled or is low power consumption when the GPUs low power consumption is well cooled the fan on the card is following that behavior it's not aware of the GDD r6 thermals at least on the nvidia cards there's no they don't program it that

Way the fan is programmed to follow the GPU temperature and so because of that if you have a GP that's particularly cool but you have a heavy memory workload and in a warm case you're going to end up with a situation where the

Fans not spinning quite fast enough to cool the memory adequately now in this card truthfully it probably does perform within spec but it's just not not the kind of thing that we're really comfortable with seeing a complete down

Costing of and the very least connects the memory to the rest of the cooler that's used for the GPU it will quote unquote increase the GPU temperature because you're sharing that cold blade that cooling solution with more

Components that are higher power consumption obviously total going into the coal blade that is but it'll be better for the overall system and a bit more metal so I mean what are you talking 5 dollars to the end user it

Just it seems like the kind of thing that shouldn't really be left out but anyway that's a complaint about the card g6 has improved over g5 that's all you need to know and now you know the exact degree to which it is improved thanks

For watching subscribe for more go to store documents next to Scott net to help side directly or patreon.com slash gamers Nexus and we'll see you all next time

Related Videos

because the Atari VCS is definitely something that's coming out we need to do a teardown on one to get some perspective and Atari if you don't know it&...
Hey how's it going guys Jack I'm out here with the toasty rose and today we're going to be doing a four hundred and fifty dollar PC that any of you ...
The biggest rule in testing coolers is to never trust anything don't trust the numbers don't trust the software don't trust the firmware and definit...
Hey what's up guys Jack and Matt here with the toaster brothers and today we're gonna be finding out if this 350 dollar laptop can game this is gonna be...
Everyone welcome back to another episode of ask GN sorry I'm not doing the intro with snowflake for this one we're flying to Taiwan and we need to get t...
Everyone welcome back to another hardware news recap for the week we are in Taipei right now working on a factory tour series and we actually have a point for y...
Hey how's it going guys Jack and Matt here with the toaster bros and today we're gonna be doing our first 20 20 thousand dollar gaming PC but guess what...
Coursers a 500 has been an embarrassing show of performance for the company's first revisit to air cooler isn't a long time but the upside is that it ma...
When I received the thread Ripper 3990 X on loan from a yet unnamed youtuber we quickly threw together an extreme overclocking stream and got the CPU to about f...
Every one so we're in Taiwan right now we're doing our factory tour series and while we were out here we stopped by Lian Lee's offices and they want...
Hey what is up guys Jack and Matt here with the wait what do you do it on sex you back your desk back to your desk sorry about that okay hi Jack and Matt here w...
We thought we'd found the limits of viewer interest when he posted our tour of a screw factory last year as poor Freight it was more of a screw shop than a ...
Everyone welcome back to another hardware news recap for the week we're in Taiwan where we venture outside to film things that's always a nice change of...
Hey how's it going guys Jack and Matt here with the toasty rose and today we're gonna be doing a 350 dollar gaming PC build it's gonna be really awe...
out of all the factory tours we've done until today we'd never gotten an opportunity to see how water blocks and open-loop cooling components like fitti...
Paints is one of the chief processes to dial in for product manufacturing and today's tour will bring us to a taiwanese factory replete with a mix of automa...
hey what's up guys Jack I'm Matt here with the tasty bros and today we're gonna be showing you our 2020 CES recap live from our CES Airbnb but befor...
except you can't wipe your ass with it well I mean we're going to take a gamble on the YouTube D monetization system and talk about the impact of corona...
everyone we're closing out our trip in Taipei for our factory tourists we have a lot of factory footage live already and we have more coming to the the chan...
hey what's up guys Jack and Matt here with the Thai stereos and today we're gonna be finding out is a six hundred dollar laptop worth it in 2020 it'...