Specific numbers are debatable, but extrapolating based on percentages is reasonable, IMO. Still, let's start by defining terms. EX: I am not going to call the 35+ million deaths from China's Great Famine as religious in nature, though some would consider the communist ideology as 'religion' as it included a state religious stance. However, the potato famine did have a very strong religious undercurrent. Clearly every death was not from famine but adding a religious multiplier seems reasonable IMO.

However, it's not just direct deaths, killing off cats during the black death significantly increased the death total resulting in millions of additional dead. Bad religious based medical care is clearly related. Blocking stem cell research is just the most recent example of superstition slowing progress, though that's much harder to estimate. More recently it's probably one of the largest killers, but I don't think this extrapolates as well to the past so I am willing to ignore it.

As to violence, historically it was much more common. Europe’s murder rate was 30 times higher in the Middle Ages than it is today. And looking at remote tribes the rates where often 20-30% of the population dead by human hands. http://reason.com/archives/2012/01/11/the-decline-of-violenc... So, religious deaths where likely more common historically, however I am willing to call this a wash.

"Cultural" practices and religions are related at some level. I would call most honor killings ~(5,000 - 20,000)/year are religious in nature. Suicides are harder to qualify, while some really are religious you can only really get at this from aggregate statistics. Indirect deaths from religious limitations on diet for example are even harder to quantify.

Religious ceremonies can also be deadly. Though again, hard to get real numbers. But, the aztecs empire killed millions (~20,000 per year), and human sacrifice was surprisingly common worldwide.

Ethic cleansing is often religious in nature specific cases are debatable, but ~100 million is probably in the right ballpark over the last 1,000 years.

Anyway, adding these up I get ~250 to 500 million over the last 1,000 years. Now find ratio from total people alive since year 1,000, subtract out the currently living population, then * ~100 billion dead.

End result I get billions, however this has a lot of assumptions and I may be underestimating these numbers in many places. For example, the 100 billion humans is based on the first human showing up 50k years ago, 200,000 years is also a reasonable number for the first human.