Razer HyperPolling Wireless Dongle

[pzogel]

Usable with any compatible Razer wireless mouse, the HyperPolling Wireless Dongle for the first time ever enables true 4000 Hz polling for a wireless gaming mouse. In this review, we'll explore and explain how the dongle works and test what benefits it offers.

Show full review

 

[OolongGoat]

I wonder if this would also apply to their wireless keyboards that use similar wireless tech?? My blackwidow mini and Naga Pro both connect to either of their respective dongles.

 

[pzogel]

I wonder if this would also apply to their wireless keyboards that use similar wireless tech?? My blackwidow mini and Naga Pro both connect to either of their respective dongles.

Mice only currently.

 

[FreedomEclipse]

We've had super high refresh rate monitors. Now its time for super high polling rate peripherals.

 

[wheresmycar]

someone enlighten me. I understand what the polling rate is in measure of (just googled it lol) but does it really help? I understand 4000Mhz means the mouse reports its position 4000 times a second... exactly where does this benefit users? Any specific workloads? My current gaming mouse is 1000Mhz and everything seems pretty instant to me. What am I missing.

 

[MrDweezil]

someone enlighten me. I understand what the polling rate is in measure of (just googled it lol) but does it really help? I understand 4000Mhz means the mouse reports its position 4000 times a second... exactly where does this benefit users? Any specific workloads? My current gaming mouse is 1000Mhz and everything seems pretty instant to me. What am I missing.

You're missing that gamers like bigger numbers, and claiming they can tell the difference between the bigger numbers and the smaller numbers.

 

[dgianstefani]

Higher frequencies improve experience when executed well, as the article goes into great detail - this isn't opinion, with the right system and screen, this hardware is objectively better than previous wireless/wired hardware with the exception of the viper 8k.

Whether it's noticeably better depends on the user and the rest of the system, but I certainly want to try this out as an upgrade to my viper.

 

[SOAREVERSOR]

We've had super high refresh rate monitors. Now its time for super high polling rate peripherals.

We've had those for a while, longer than the monitors actually.

 

[Dr. Dro]

All things considered, this is probably the weirdest gimmick I've ever seen, anyone who knows half a thing about input polling will run away from this device like a vampire runs away from garlic and the cross. More Hz does not necessarily mean better in this case.

If you value performance and need a little extra accuracy, stick to 250 Hz. For keyboards, even 125 Hz should be fine for most cases.

 

[gffermari]

So, you buy a 1000Hz mice and with that you convert it to a 4000Hz.
Why don't you buy a 4/8KHz from the beginning?

It only supports the latest razer products, Razer Basilisk V3 Pro, Razer DeathAdder V3 Pro, Razer Viper V2 Pro....

 

[Dr. Dro]

So, you buy a 1000Hz mice and with that you convert it to a 4000Hz.
Why don't you buy a 4/8KHz from the beginning?

It only supports the latest razer products, Razer Basilisk V3 Pro, Razer DeathAdder V3 Pro, Razer Viper V2 Pro....

There is a good reason that 99.99% of mice don't offer polling rates beyond 1000 Hz over USB. You shouldn't run the polling rate that high to begin with. For most high performance USB optical and laser mice you want a polling rate of 250 Hz at most. Keyboards can run at the standard 125 Hz polling rate perfectly well unless you reliably exceed 125 wps (yes, second, not minute, at that point, where did you get the Ghost in the Shell doctor's hand augmentation, because I want it, it'd come in handy in more than a few ways).

Hardware interrupts are something that you want to keep at a minimum at all times, the benefits of doing this far exceed anything you could possibly get from a 4000 Hz polling rate. There is a high likelihood that software will misbehave when trying to poll raw data from such a high-frequency device, and this will come at a very high CPU usage cost, especially over USB. It seems to strain even the mice's own signal processing chip given the drastic decrease in battery life. This device is a gimmick, and while I wouldn't call it snake oil (as it delivers what it promises), it's about as unwise to do so as it comes, even if you have a state of the art 360 Hz monitor or an OLED panel.

 

[dgianstefani]

So, you buy a 1000Hz mice and with that you convert it to a 4000Hz.
Why don't you buy a 4/8KHz from the beginning?

It only supports the latest razer products, Razer Basilisk V3 Pro, Razer DeathAdder V3 Pro, Razer Viper V2 Pro....

The 8KHz was in wired only, since physics means the battery drain would be too significant for wireless.

The 4KHz/2KHz hyperpolling unit wasn't ready when the Viper V2 Pro launched - you get 50% off if you use the 4000hz code when purchasing the hyperpolling module.

Of course it only supports the latest products, it's a new standard?

There is a good reason that 99.99% of mice don't offer polling rates beyond 1000 Hz over USB. You shouldn't run the polling rate that high to begin with. For most high performance USB optical and laser mice you want a polling rate of 250 Hz at most. Keyboards can run at the standard 125 Hz polling rate perfectly well unless you reliably exceed 125 wps (yes, second, not minute, at that point, where did you get the Ghost in the Shell doctor's hand augmentation, because I want it, it'd come in handy in more than a few ways).

Hardware interrupts are something that you want to keep at a minimum at all times, the benefits of doing this far exceed anything you could possibly get from a 4000 Hz polling rate. There is a high likelihood that software will misbehave when trying to poll raw data from such a high-frequency device, and this will come at a very high CPU usage cost, especially over USB. It seems to strain even the mice's own signal processing chip given the drastic decrease in battery life. This device is a gimmick, and while I wouldn't call it snake oil (as it delivers what it promises), it's about as unwise to do so as it comes, even if you have a state of the art 360 Hz monitor or an OLED panel.

This isn't the 1990s, or the 2000s, or the 2010s, we have CPUs that have excess of 8 cores, with very high clockspeed and large caches, combined with high IPC architectures, and interconnect subsystems that operate at exponentially faster speeds than previously. The kind of prosumer/gamer who buys this product won't be someone using a 10 year old i5, or an AMD FX CPU.

Your comments are true, to an extent, but your conclusions are wrong. The "high" CPU cost is only relative to 1KHz or lower polling hardware, it's still nothing that will overwhelm a well built, modern PC, and they're only going to get faster. Your presumption that the battery life is reduced because the "signal processing chip" is "strained" is just an assumption, and probably wrong - the battery life is affected because the module is transmitting 2x or 4x the data, with no changes to battery capacity. Basic physics.

If noone pushes the envelope, hardware doesn't get better and more refined. 10 years from now there'll be modules that combine high frequency with high density batteries, having no compromises.

Your argument based around "anything higher than 125Hz is pointless because you can't type more than 125wpm" is fundamentally flawed in the context of the product and it's target group: gamers and typists have different needs, a typist cares little for the latency between their keypress and the word, because they're doing a sequential, predictable task. A gamer wants to eliminate as much latency between their key or button press, and results on the screen, this can be attacked by improving internet connection, reducing processing time, increasing frequency of data updates etc.

 

[Dr. Dro]

This isn't the 1990s, or the 2000s, or the 2010s, we have CPUs that have excess of 8 cores, with very high clockspeed and large caches, combined with high IPC architectures, and interconnect subsystems that operate at exponentially faster speeds than previously. The kind of prosumer/gamer who buys this product won't be someone using a 10 year old i5, or an AMD FX CPU.

Your comments are true, to an extent, but your conclusions are wrong. The "high" CPU cost is only relative to 1KHz or lower polling hardware, it's still nothing that will overwhelm a well built, modern PC, and they're only going to get faster. Your presumption that the battery life is reduced because the "signal processing chip" is "strained" is just an assumption, and probably wrong - the battery life is affected because the module is transmitting 2x or 4x the data, with no changes to battery capacity. Basic physics.

If noone pushes the envelope, hardware doesn't get better and more refined. 10 years from now there'll be modules that combine high frequency with high density batteries, having no compromises.

Your argument based around "anything higher than 125Hz is pointless because you can't type more than 125wpm" is fundamentally flawed in the context of the product and it's target group: gamers and typists have different needs, a typist cares little for the latency between their keypress and the word, because they're doing a sequential, predictable task. A gamer wants to eliminate as much latency between their key or button press, and results on the screen, this can be attacked by improving internet connection, reducing processing time, increasing frequency of data updates etc.

The assumption that the reduced battery life comes from data transfer rate counts on the RF chipset increasing the transmit power significantly, which is quite unlikely - the increased processing power required by the faster rate of polling has a far higher likelihood of being the cause of the increased power consumption. Anyway - that's speculation on both of our parts, only someone who knows how the product's RF system actually works can answer definitively.

The thing is that it doesn't matter how fast PCs get, or whether the interrupt request comes from a mouse or a keyboard, an excessive rate will always be detrimental to system performance - it is a basic architecture quirk because PCs are still designed to be first and foremost compatible with the original PC architecture, even today. Whether you are moving your mouse cursor or typing a key on your keyboard, this is achieved by sending a signal interrupt from the device to the host processor, and this is actually one of the fundamentals of operating system programming. You can quickly Google "mouse polling rate performance issues" and get a lot of answers from people advising others to lower from 1000 to say, 500 Hz to eliminate hitching, that would be the reason why.

You'll find the impact of polling rates to be actually benchmarkable through many means, be it with your favorite frame profiler (such as CapframeX) or by using LatencyMon to measure ISRs/DPC latency in general. If a gamer wants to eliminate as much latency between their key and button press, I could write in a lot more detail but i'll just keep it short, I have a proposition: quit online games entirely. Ping is bad mmkay, and it's not this potential 0.05 ms at the cost of flooding your OS with an extra 3000-4000 interrupts a second (ISRs to the moon!) that's going to change the outcome of an online match, just the variable network latency between clients is a wildcard beyond solving at present. I can see this being moderately useful to people who play games like osu! but at the same time, such extreme polling rates may very well induce systems to hitching, which would be far worse than an physiologically imperceptible increase (to most, anyway) in response time.

 

[dgianstefani]

The assumption that the reduced battery life comes from data transfer rate counts on the RF chipset increasing the transmit power significantly, which is quite unlikely - the increased processing power required by the faster rate of polling has a far higher likelihood of being the cause of the increased power consumption. Anyway - that's speculation on both of our parts, only someone who knows how the product's RF system actually works can answer definitively.

The thing is that it doesn't matter how fast PCs get, or whether the interrupt request comes from a mouse or a keyboard, an excessive rate will always be detrimental to system performance - it is a basic architecture quirk because PCs are still designed to be first and foremost compatible with the original PC architecture, even today. Whether you are moving your mouse cursor or typing a key on your keyboard, this is achieved by sending a signal interrupt from the device to the host processor, and this is actually one of the fundamentals of operating system programming. You can quickly Google "mouse polling rate performance issues" and get a lot of answers from people advising others to lower from 1000 to say, 500 Hz to eliminate hitching, that would be the reason why.

You'll find the impact of polling rates to be actually benchmarkable through many means, be it with your favorite frame profiler (such as CapframeX) or by using LatencyMon to measure ISRs/DPC latency in general. If a gamer wants to eliminate as much latency between their key and button press, I could write in a lot more detail but i'll just keep it short, I have a proposition: quit online games entirely. Ping is bad mmkay, and it's not this potential 0.05 ms at the cost of flooding your OS with an extra 3000-4000 interrupts a second (ISRs to the moon!) that's going to change the outcome of an online match, just the variable network latency between clients is a wildcard beyond solving at present. I can see this being moderately useful to people who play games like osu! but at the same time, such extreme polling rates may very well induce systems to hitching, which would be far worse than an physiologically imperceptible increase (to most, anyway) in response time.

Again, some salient points, but your statistics are off by a factor of 10. The difference is 0.5ms, not 0.05ms, on average moving from 1KHz to 4KHz polling, but can vary between 0.25 and 0.75ms. I'm going to have a nap now as i've been proofreading all morning but it's just one hole I can point at.

You would know this if you read the review, the relevant data is helpfully graphed.

Again, if your system is too slow to take advantage of this technology without hitching, then you probably aren't the target audience.

No need to take personal offence at products aimed higher than what you are happy with.

This is click latency as measured from the button press to the pixel changing, so it takes into account any "performance issues" that the increased polling incurs, which seem to be irrelevant on the system tested.

 

[Dr. Dro]

Again, some salient points, but your statistics are off by a factor of 10. The difference is 0.5ms, not 0.05ms, on average moving from 1KHz to 4KHz polling, but can vary between 0.25 and 0.75ms. I'm going to have a nap now as i've been proofreading all morning but it's just one hole I can point at.

You would know this if you read the review, the relevant data is helpfully graphed.

View attachment 259560

Again, if your system is too slow to take advantage of this technology without hitching, then you probably aren't the target audience.

No need to take personal offence at products aimed higher than what you are happy with.

Mmm, yes, it's 0.5, not 0.05, my bad. But it still does not change much of my reasoning.

I did read the review, and as you can see, my rig is plenty fast. I still wouldn't use this. There are too many caveats and potential downsides, as well as complex technical reasons as to why a higher polling rate is not necessarily a beneficial thing. My "explanation" is also greatly oversimplified, for example, I more or less folded the whole "USB host controller interfacing" thing directly, for example.

Anyway, cheers on you getting that job mate. It's a good one.

 

[dgianstefani]

Mmm, yes, it's 0.5, not 0.05, my bad. But it still does not change much of my reasoning.

I did read the review, and as you can see, my rig is plenty fast. I still wouldn't use this. There are too many caveats and potential downsides, as well as complex technical reasons as to why a higher polling rate is not necessarily a beneficial thing. My "explanation" is also greatly oversimplified, for example, I more or less folded the whole "USB host controller interfacing" thing directly, for example.

Anyway, cheers on you getting that job mate. It's a good one.

Thanks. Always good to have intelligent discussion without name-calling.

OptimumTech has good things to say.

 

[R-T-B]

125 Hz polling rate perfectly well unless you reliably exceed 125 wps (yes, second, not minute, at that point, where did you get the Ghost in the Shell doctor's hand augmentation, because I want it, it'd come in handy in more than a few ways).

It's also characters, not words.

My mom got hers at Court Reporting school.

Hardware interrupts

Aren't used in polled hardware, like USB. There is some argument for keeping the polling rate down, but it has nothing to do with hardware interrupts.

 

[Dr. Dro]

It's also characters, not words.

My mom got hers at Court Reporting school.


Aren't used in polled hardware, like USB. There is some argument for keeping the polling rate down, but it has nothing to do with hardware interrupts.

May have been a bit tired when I wrote all that, so take it with bag of salt

But yea, interrupts and polling more or less just change the side that is being requested, so in both cases having excessive rates of polling or interrupts will degrade performance